If it happens after January I can lend a helping hand and do a PCB in Eagle and send you the Gerber files. Send me an email if you need some help on that. I will be gone in January.
Interest in very simple but usable oscilloscope?
- Thread starter womai
- Start date
Hi RexLan,
thanks for the offer! But actually the PCB will be the smallest part of the work - I can re-use portions of the existing DPScope design, and the circuit is relatively simple (few components), so it will be just a day or two to whip this up. Wouldn't make much sense to re-do this in Eagle. The major effort is the PIC firmware, followed by the necessary changes to the PC software. And yes, "after January" is a good bet - right now I am super-busy with my real job.
Regards,
Wolfgang
thanks for the offer! But actually the PCB will be the smallest part of the work - I can re-use portions of the existing DPScope design, and the circuit is relatively simple (few components), so it will be just a day or two to whip this up. Wouldn't make much sense to re-do this in Eagle. The major effort is the PIC firmware, followed by the necessary changes to the PC software. And yes, "after January" is a good bet - right now I am super-busy with my real job.
Regards,
Wolfgang
My condolences ... that really sucks!
Here I thought I was being punished having to go to Florida for a month fishing ... I feel much better now.
There are a lot of design choices and tradeoffs when designing an oscilloscope (like for any type of circuit). Cost, ease of assembly, performance, firmware development and so on.
This oscilloscope is above all meant to be very(!) inexpensive, and also simple to build, hence the PIC (18FR4550 or 18F2550) is a pretty good choice - integrates controller, capture memory, ADCs, PWM (e.g. for trigger level), USB interface all in one cheap, easy-to-solder DIP package. The rest of the circuit is then just a few components. Plus I think I will be able to add a simple logic analyzer (i.e. 8 digital input channels) and a very simple arbitrary waveform generator (or 8 digital output channels using the same PIC outputs) almost for free. So a very comprehensive instrument for very little expense. Of course the main tradeoff here is performance - but it should be good enough for 90% of what a typical Picaxe user does.
For quite a bit more power you can go e.g. to a dsPIC - but wait, I've already done that
(see http://www.dpscope.com). Downside - to actually use the performance you need a better frontend, the USB interface is no longer integrated (so additional cost), and so on, all that increases cost - and you don't get logic analyzer or AWG functionality on the DPScope - but you get about 10x the sample rate and bandwidth. So it's a somewhat different tradeoff.
Microcontroller based solutions quickly reach their limit when you use them to perform things like sampling control in an oscilloscope. So even the fastest ones (something like a 100 MHz 32 bit model) will only give underwhelming performance (a few MSamples/sec at best) when compared to a dedicated logic solution. If I wanted to create a higher end instrument than that I think the best approach is to use an FPGA for sample logic and control - even small FPGAs nowadays, like the Xilinx Spartan series - have enough power to achieve ~100-200 MSamples/sec (good enough for 10-20 MHz analog bandwidth in single shot mode), and can integrate a microcontroller softcore like e.g. the Microblaze. But such an instrument will cost at least ~$200-$300 to build in a small series since to make use of the FPGAs performance you need better amplifiers, fast external ADCs, external USB interface and so on. So in this case it's probably better to just go out and buy a Bitscope, Velleman or similar instrument which sells for about what a self-built one would cost, if not less.
So to sum up, the reason for this exercise is really to see how low-cost I can go and still produce something that is practically useful.
Wolfgang
This oscilloscope is above all meant to be very(!) inexpensive, and also simple to build, hence the PIC (18FR4550 or 18F2550) is a pretty good choice - integrates controller, capture memory, ADCs, PWM (e.g. for trigger level), USB interface all in one cheap, easy-to-solder DIP package. The rest of the circuit is then just a few components. Plus I think I will be able to add a simple logic analyzer (i.e. 8 digital input channels) and a very simple arbitrary waveform generator (or 8 digital output channels using the same PIC outputs) almost for free. So a very comprehensive instrument for very little expense. Of course the main tradeoff here is performance - but it should be good enough for 90% of what a typical Picaxe user does.
For quite a bit more power you can go e.g. to a dsPIC - but wait, I've already done that
(see http://www.dpscope.com). Downside - to actually use the performance you need a better frontend, the USB interface is no longer integrated (so additional cost), and so on, all that increases cost - and you don't get logic analyzer or AWG functionality on the DPScope - but you get about 10x the sample rate and bandwidth. So it's a somewhat different tradeoff.Microcontroller based solutions quickly reach their limit when you use them to perform things like sampling control in an oscilloscope. So even the fastest ones (something like a 100 MHz 32 bit model) will only give underwhelming performance (a few MSamples/sec at best) when compared to a dedicated logic solution. If I wanted to create a higher end instrument than that I think the best approach is to use an FPGA for sample logic and control - even small FPGAs nowadays, like the Xilinx Spartan series - have enough power to achieve ~100-200 MSamples/sec (good enough for 10-20 MHz analog bandwidth in single shot mode), and can integrate a microcontroller softcore like e.g. the Microblaze. But such an instrument will cost at least ~$200-$300 to build in a small series since to make use of the FPGAs performance you need better amplifiers, fast external ADCs, external USB interface and so on. So in this case it's probably better to just go out and buy a Bitscope, Velleman or similar instrument which sells for about what a self-built one would cost, if not less.
So to sum up, the reason for this exercise is really to see how low-cost I can go and still produce something that is practically useful.
Wolfgang
Last edited:
I'm starting to fear I should have made clear I meant January 2012 All of last week was spent on a business trip, so you can imagine how much progress this means. I do keep juggling the design ideas for the scope in my head though. Hate to make everybody (including myself) wait, but I guess that's life. At least I still find a few minutes to read this forum...
Wolfgang
All of last week was spent on a business trip, so you can imagine how much progress this means. I do keep juggling the design ideas for the scope in my head though. Hate to make everybody (including myself) wait, but I guess that's life. At least I still find a few minutes to read this forum...Wolfgang
Good things take time, womai. We'll wait, quietly, but eagerly.
(Note that at the time of this post there are 2,394 views of this thread, so there's a bit of interest.)
Sorry to hear you are working so hard, but I'm pleased to hear that at least someone is working a full-time job.
(Note that at the time of this post there are 2,394 views of this thread, so there's a bit of interest.)
Sorry to hear you are working so hard, but I'm pleased to hear that at least someone is working a full-time job.
Last edited:
I'm very interested in this; I need an oscilloscope for a personal project I'm working on but I'm living in China at the moment making $600 a month so anything to save some cash comes in handy; this DIY one seems to be exactly what I need but I'd rather wait for the updated version rather than use the old one (although it's still quite functional!)
Any updates the past few weeks?
Any updates the past few weeks?
Hi, with "the old one" do you mean my DPScope (www.dpscope.com)? That one and the planned one are difficult to compare - the DPScope has much better performance as a scope (10x higher bandwidth, 10x higher sample rate), the planned one has additional functionality (waveform generator, and potentially logic analyzer and digital pattern generator, albeit don't expect very high data rates for either. Keep in mind though, for now the new scope is a pure paper scope, and with my current workload I have no idea if/when I will be able to get back to it. If you want, I also sell just the pre-programmed microcontroller and printed circuit board for the DPScope - that may be a good deal if you can get the other components cheaply or have many already lying around in your drawer.
Wolfgang
Wolfgang
Hehe, sorry, I meant this one which I believe was completed by you right? If not I suppose the one from the first post was the only one completed, is that the case? In any event, I don't need anything fancy. I'm working on a 22kHz ultrasonic transmitter for underwater communication and just need to check that my components are working properly before I get too deep into the project only to find that it wasn't feasible to begin with. Speaking of which you mentioned including a hex file for the one on the first page but I didn't see it in this thread? Do you still have it?
As much as I'd like to buy the DPScope or the preprogrammed chips, it wouldn't make much logical sense for me to do so. I'm teaching in China at the moment (adventure, learn the language, blah blah) and make $600 a month. That's good pay in China but once my student loans take their piece out of me it leaves little so I must spend wisely. For now, I need to get a PIC programmer, tools (soldering iron, multimeter, breadboard, wires, etc), and rudimentary electronic equipment so that I can begin testing transducers, MCUs, and voltages. I'll buy more as I go along. This puts a low cost oscilloscope high on my list of things to procure.
Interesting note: some electronic parts are cheap here but you'd be surprised that things usually cost the same or more (especially computers!). After all, why sell at a cheap price in China when they can sell it for much more by exporting? Supply and demand in action. Fortunately, components are (usually) cheap worldwide.
As much as I'd like to buy the DPScope or the preprogrammed chips, it wouldn't make much logical sense for me to do so. I'm teaching in China at the moment (adventure, learn the language, blah blah) and make $600 a month. That's good pay in China but once my student loans take their piece out of me it leaves little so I must spend wisely. For now, I need to get a PIC programmer, tools (soldering iron, multimeter, breadboard, wires, etc), and rudimentary electronic equipment so that I can begin testing transducers, MCUs, and voltages. I'll buy more as I go along. This puts a low cost oscilloscope high on my list of things to procure.
Interesting note: some electronic parts are cheap here but you'd be surprised that things usually cost the same or more (especially computers!). After all, why sell at a cheap price in China when they can sell it for much more by exporting? Supply and demand in action. Fortunately, components are (usually) cheap worldwide.
Womai has made quite an amazing, low cost 'scope there. However, if you can't spend that much, I found this review on tronixstuff quite interesting. Small and cheap scope, even if it's not really high powered in any sense.
Yes, that's what impresses me most about Womai's is that it can get so much functionality at such a cost. Sure it's not meant to be professional grade but dang it does more than I need it to! Honestly, his isn't expensive it's simply the comparative costs that prohibit me from doing so (e.g. I am paid in RMB so comparatively the shipping for the DPScope kit costs me the equivalent of roughly $140).
If I were to go with a standalone scope, I'd probably go with the DSO Nano which can still be found on Ebay for around $45 (or on some online retailers around $60 but getting harder to find since they came out with V2 which is $90). Still, I have such a large screen on my laptop, it would be much better to have a cheaper one to just plug in
If I were to go with a standalone scope, I'd probably go with the DSO Nano which can still be found on Ebay for around $45 (or on some online retailers around $60 but getting harder to find since they came out with V2 which is $90). Still, I have such a large screen on my laptop, it would be much better to have a cheaper one to just plug in
I breadboarded the 16F684-based scope (minus the analog frontend), but the software (firmware and PC software) is still in the proof-of-concept stage. Once I was that far I figured I could use the 18F4550 instead which has USB built in, so for a little bit more money you'd get a USB interface instead of the serial one, could thus power the scope through USB (saves the cost of an external power supply), get single-shot capture to at least 50 KSa/sec, and add a lot of additional functionality. So chances for the 16F684-scope to be ever completed are approximately zero. If you want, feel free to use the schematic and build your own scope based on it. You'll need to write your own software and firmware though...
I would not recommend the DSO Nano - there have been other threads discussing its merits, but it only has a single channel, which makes it useless for many applications where you need to look at one signal relative to another, or trigger on a signal other than the one you want to look at. So its more an admittedly cute toy than a real instrument.
I would not recommend the DSO Nano - there have been other threads discussing its merits, but it only has a single channel, which makes it useless for many applications where you need to look at one signal relative to another, or trigger on a signal other than the one you want to look at. So its more an admittedly cute toy than a real instrument.
Ah, that's a shame. Based on what I need it for now, perhaps I'll just wait to get the scope later. Really I just needed it to measure signal strength of the receiving transducer and I guess I can accomplish that by sticking an LED on the receiver and seeing how far I can get away at various voltages or using different transducers.
I really like the idea, though! Hope you finish it someday.
I really like the idea, though! Hope you finish it someday.
Made some progress - attached is the most recent schematic. Not breadboarded or tested in practice yet, so if a anybody spots anything funny, let me know.
I tried to put in a lot of functionality without adding much cost. On top of that, the circuit is pretty modular, so e.g. if you have no need for an AWG (arbitrary waveform generator), just leave out the components for the R/2R network and the low-pass reconstruction filter. Or if you don't need BNC connectors for scope probes, don't install them (scope channels come out on jumper header on back in addition, good for breadboarding).
Here are the main specs:
Global:
- based on Microchip PIC18F2550
- display and control on PC
- data and power through USB
- 2 BNC connectors on frontpanel
- 0.1" header (2 x 10 pins) on back
- will fit into optional enclosure (Serpac A-21, like DPScope) with custom endpanels
Scope:
- two channels, 1 MOhm input impedance, +/-25V range
- trigger on CH1, or free run; rising or falling edge, adjustable threshold
- real-time sampling rate 50 kSa/sec on two channels (100 kSa/sec on one)
- equivalent time sampling rate (for repetitive signals) probably up to 2 MSa/sec
- analog bandwidth around 200 kHz (not tested yet)
- record length 256 samples per channel
- FFT (done on PC)
Logic analyzer (LA):
- 4 channels, 10 kOhm input impedance, 5V TTL or CMOS logic levels
(4 channels is good enough for serial busses like RS-232, I2C, SPI, CAN)
- real-time sampling rate 100 kSa/sec (maybe a bit more)
- record length ~1500-2000 samples (hopefully) if no scope channels active
- hardware trigger on 2-state sequence (may change)
Digital pattern generator (PG):
- 8 digital channels
- max. bit rate 100 kHz, adjustable
- 5V logic
Arbitrary waveform generator (AWG):
- 1 channel, range 0-5V (DC coupled) or AC coupled (5 V max amplitude)
- DDS-based, i.e. fine frequency resolution (sub-Hz), sample rate 100 kHz
- waveform memory 256 samples
- DAC resolution 8 bits; low-pass reconstruction filter
- analog bandwidth 20 kHz
PG and AWG are mutually exclusive since they use the same microcontroller pins. Scope and LA may be mutually exclusive (to maximize record length for LA).
Wolfgang
I tried to put in a lot of functionality without adding much cost. On top of that, the circuit is pretty modular, so e.g. if you have no need for an AWG (arbitrary waveform generator), just leave out the components for the R/2R network and the low-pass reconstruction filter. Or if you don't need BNC connectors for scope probes, don't install them (scope channels come out on jumper header on back in addition, good for breadboarding).
Here are the main specs:
Global:
- based on Microchip PIC18F2550
- display and control on PC
- data and power through USB
- 2 BNC connectors on frontpanel
- 0.1" header (2 x 10 pins) on back
- will fit into optional enclosure (Serpac A-21, like DPScope) with custom endpanels
Scope:
- two channels, 1 MOhm input impedance, +/-25V range
- trigger on CH1, or free run; rising or falling edge, adjustable threshold
- real-time sampling rate 50 kSa/sec on two channels (100 kSa/sec on one)
- equivalent time sampling rate (for repetitive signals) probably up to 2 MSa/sec
- analog bandwidth around 200 kHz (not tested yet)
- record length 256 samples per channel
- FFT (done on PC)
Logic analyzer (LA):
- 4 channels, 10 kOhm input impedance, 5V TTL or CMOS logic levels
(4 channels is good enough for serial busses like RS-232, I2C, SPI, CAN)
- real-time sampling rate 100 kSa/sec (maybe a bit more)
- record length ~1500-2000 samples (hopefully) if no scope channels active
- hardware trigger on 2-state sequence (may change)
Digital pattern generator (PG):
- 8 digital channels
- max. bit rate 100 kHz, adjustable
- 5V logic
Arbitrary waveform generator (AWG):
- 1 channel, range 0-5V (DC coupled) or AC coupled (5 V max amplitude)
- DDS-based, i.e. fine frequency resolution (sub-Hz), sample rate 100 kHz
- waveform memory 256 samples
- DAC resolution 8 bits; low-pass reconstruction filter
- analog bandwidth 20 kHz
PG and AWG are mutually exclusive since they use the same microcontroller pins. Scope and LA may be mutually exclusive (to maximize record length for LA).
Wolfgang
Attachments
Last edited:
hippy
Ex-Staff (retired)
It may just be my own personal preference, but I like to put CH1 as the 'main trace' and trigger on CH2, usually a pulse generated just before the data I want to look at.
Had a little fun putting a Spice model of the input stage together. Encouraging news - that portion works just as intended (at least in simulation). Analog bandwidth is 260 kHz for the gain 10 path (+/-2.5V input signal range), and close to 300 kHz for the gain 1 path (+/-25V range), so my initial guess (200 kHz) wasn't far off. Of course it now depends what the bandwidth of the PIC's ADC is...
I used Microchip's free Mindi simulator for this work, so if anybody is interested, feel free to play with it as well. Sees like Microchip is no longer supporting the downloadable, standalone Mindi (luckily I was able to still find it on their website using the search tool) and moved to a dumbed-down web-based interface
Attached is the schematic and the Bode (gain vs. frequency) plot.
Wolfgang
I used Microchip's free Mindi simulator for this work, so if anybody is interested, feel free to play with it as well. Sees like Microchip is no longer supporting the downloadable, standalone Mindi (luckily I was able to still find it on their website using the search tool) and moved to a dumbed-down web-based interface
Attached is the schematic and the Bode (gain vs. frequency) plot.
Wolfgang
Attachments
Very nice, for most of us here, would be a very usable piece of equipment.
I like the use of an 18F2550, I have a few 18F4550's on hand when I was playing around with USB connectivity, it wouldn't be hard to fit in on a 28 pin board... less cost to me...
I already have your DPScope and it has done what I asked of it, I know there is an upgraded chip for it, but I just don't need it... yet...
You mentioned about the dspic30F2020 used in the DPScope and the possibility of upgrading the chip, I gather it is this one... http://ww1.microchip.com/downloads/en/DeviceDoc/70343A.pdf it might be 5volt tolerant but the DAC seems to be 3.3volt.
Cheers and keep up the good work Wolfgang, I don't mind waiting too, as I know a 'hobby' takes a back seat to a real job and raising a family
I like the use of an 18F2550, I have a few 18F4550's on hand when I was playing around with USB connectivity, it wouldn't be hard to fit in on a 28 pin board... less cost to me...
I already have your DPScope and it has done what I asked of it, I know there is an upgraded chip for it, but I just don't need it... yet...
You mentioned about the dspic30F2020 used in the DPScope and the possibility of upgrading the chip, I gather it is this one... http://ww1.microchip.com/downloads/en/DeviceDoc/70343A.pdf it might be 5volt tolerant but the DAC seems to be 3.3volt.
Cheers and keep up the good work Wolfgang, I don't mind waiting too, as I know a 'hobby' takes a back seat to a real job and raising a family
megareg,
thanks for the praise
yes, the 18F2550/4550 is an amazing piece of hardware - once you finally have the clock tree set up correctly and the PC side of the USB connection figured out, that is
I have been closely following the new dsPIC33F chip you mention. It would require some hardware and software changes since it runs at 3.3V, not 5V like the dsPIC30F2020. On the upside that would mean I could use a simple low-dropout linear voltage regulator to clean up the sometimes noisy USB supply voltage. It would also give more RAM for longer waveform records, and allow for double the sample rate (although I am not yet sure the chip itself runs fast enough to really achieve that - the 2020 right now uses 100% of its computing power to capture and store at 1 MSa/sec on two channels...).
But unfortunately the Errata sheet for this chip reveals a real showstopper - triggering on the comparator for a negative slope (falling edge) does not work correctly, and triggering is such a crucial function for the scope that right now I am holding off on any development work with this chip.
Wolfgang
thanks for the praise
yes, the 18F2550/4550 is an amazing piece of hardware - once you finally have the clock tree set up correctly and the PC side of the USB connection figured out, that is
I have been closely following the new dsPIC33F chip you mention. It would require some hardware and software changes since it runs at 3.3V, not 5V like the dsPIC30F2020. On the upside that would mean I could use a simple low-dropout linear voltage regulator to clean up the sometimes noisy USB supply voltage. It would also give more RAM for longer waveform records, and allow for double the sample rate (although I am not yet sure the chip itself runs fast enough to really achieve that - the 2020 right now uses 100% of its computing power to capture and store at 1 MSa/sec on two channels...).
But unfortunately the Errata sheet for this chip reveals a real showstopper - triggering on the comparator for a negative slope (falling edge) does not work correctly, and triggering is such a crucial function for the scope that right now I am holding off on any development work with this chip.
Wolfgang
Last edited:
It has been a few weeks since an update on Wolfgang's "cheap-scope" project. I certainly could have used one today to send out with my boss on an equipment maintenance trip.
One of the "undocumented" features of this "cheap-scope" project is that it will most certainly be tiny. That's a very handy feature when you're already toting laptop PC's and luggage through airport terminals, and having it all examined (abused) by suspicious TSA employees. They ignore innocuous little items, but a big O'scope makes them all get very nervous. It looks "technical," and they don't understand technical.
But if Wolfgang's as busy as I am right now, (I worked at the office till 3:00 A.M. this morning and I'm pretty weary,) then I well understand why his side projects are in limbo.
Hang in there, Womai!
One of the "undocumented" features of this "cheap-scope" project is that it will most certainly be tiny. That's a very handy feature when you're already toting laptop PC's and luggage through airport terminals, and having it all examined (abused) by suspicious TSA employees. They ignore innocuous little items, but a big O'scope makes them all get very nervous. It looks "technical," and they don't understand technical.
But if Wolfgang's as busy as I am right now, (I worked at the office till 3:00 A.M. this morning and I'm pretty weary,) then I well understand why his side projects are in limbo.
Hang in there, Womai!
Well, I actually did work a little bit on this. Was happy to discover Microchip's 20-pin USB capable 18F14K50. That thing is much smaller and cheaper than the 18F2550 so would be a great fit. Not capable enough for all functions at once, but could either work in the AWG or in the scope. And yes, that would enable a much smaller board. Got really enthusiastic, ordered a couple from Digikey, completed the schematic for the AWG and a basic version of the AWG firmware while waiting for them (AWG is much less software and firmware development so I thought I'd use that as my practice platform to get USB etc. up and running on this device).
Got them Friday, tried them today.... and had to discover that my d***d development board does not support them. (the compiler is fine with them, but the programmer refuses support). So now I am back to either using the 2550 or to shell out hard cash for a new development board. Need some time to reflect...
Got them Friday, tried them today.... and had to discover that my d***d development board does not support them. (the compiler is fine with them, but the programmer refuses support). So now I am back to either using the 2550 or to shell out hard cash for a new development board. Need some time to reflect...
Wanted to give you a little update: While waiting for the PIC18F14K50 USB development platform (supplied to me be John West - a big thank you on behalf of the whole Picaxe community!) I drew up a much simplified scope schematic. The 14K50 isn't big enough for all the features to be put in, so I'll split the design into a scope and into an AWG/digital pattern generator. Advantage - chance are one of them will be done much earlier (especially the software!) than the one-thing-does-it-all monster
Attached is the preliminary schematic. There is a grand total of two DIP-packaged chips (the microcontroller and a quad-op-amp) that do all the lifting, plus a few discrete parts - nothing exotic anywhere, and all through hole. Should make for a small board that is easy to put together. Record length will be ~200pts per channel, sample rate ~100 kSa/sec real time and maybe 1-2 MSa/sec equivalent time (for repetitive signals). All powered from the USB port. I'll make it a HID-type USB device so there won't even be a driver to install.
On the PC side I should be able to re-use a lot of the code from DPScope PC software.
Wolfgang
Attached is the preliminary schematic. There is a grand total of two DIP-packaged chips (the microcontroller and a quad-op-amp) that do all the lifting, plus a few discrete parts - nothing exotic anywhere, and all through hole. Should make for a small board that is easy to put together. Record length will be ~200pts per channel, sample rate ~100 kSa/sec real time and maybe 1-2 MSa/sec equivalent time (for repetitive signals). All powered from the USB port. I'll make it a HID-type USB device so there won't even be a driver to install.
On the PC side I should be able to re-use a lot of the code from DPScope PC software.
Wolfgang
Attachments
Some more progress. See attached schematic and layout.
I added a connector that feeds out 4 I/O pins, power and ground - this could be used as a 4-bit wide logic analyzer (LA) input. In addition it provides the 5V USB supply, this can be used to power e.g. a breadboard circuit. There isn't enough RAM or computing power to support scope and LA at the same time, but the LA on its own could sample about 800 - 1000 samples at 100-200 kSamples/sec. Should be good enough to debug RS-232, slow-speed I2C, one-wire and so on. Most of the work there will be the software.
Also added (optional) capacitive input compensation (the two capacitors in parallel to the resistive input divider) in case the true bandwidth turns out to be less than the simulation predicts. I made all the capacitors fixed value (not trimmers like in the DPScope) to keep down cost.
Also finished the first pass of the layout - by chance the board worked out to almost the same size as the DPScope board, so I made it exactly the same size, meaning it will fit in the same plastic enclosure (Serpac A-21). So I kept the custom end panels as part of the PCB - should make for a very nice-looking instrument. Decided it will be called "DPScope SE" (even though there is no longer any dsPIC inside
My preliminary guess is about US$30 to put one together including plastic enclosure.
Now waiting for the programmer and development board so I can continue working on firmware and software.
Wolfgang
I added a connector that feeds out 4 I/O pins, power and ground - this could be used as a 4-bit wide logic analyzer (LA) input. In addition it provides the 5V USB supply, this can be used to power e.g. a breadboard circuit. There isn't enough RAM or computing power to support scope and LA at the same time, but the LA on its own could sample about 800 - 1000 samples at 100-200 kSamples/sec. Should be good enough to debug RS-232, slow-speed I2C, one-wire and so on. Most of the work there will be the software.
Also added (optional) capacitive input compensation (the two capacitors in parallel to the resistive input divider) in case the true bandwidth turns out to be less than the simulation predicts. I made all the capacitors fixed value (not trimmers like in the DPScope) to keep down cost.
Also finished the first pass of the layout - by chance the board worked out to almost the same size as the DPScope board, so I made it exactly the same size, meaning it will fit in the same plastic enclosure (Serpac A-21). So I kept the custom end panels as part of the PCB - should make for a very nice-looking instrument. Decided it will be called "DPScope SE" (even though there is no longer any dsPIC inside
My preliminary guess is about US$30 to put one together including plastic enclosure.
Now waiting for the programmer and development board so I can continue working on firmware and software.
Wolfgang
Attachments
That's a good-looking layout womai. I have two comments about the USB
1) From what I've seen it's normal to include a 500mA fuse in series with the 5v
2) Do you need to do anything with D+ and D- if you aren't using them? Will the caps affect the signals?
http://www.sump.org/projects/analyzer/client/
1) From what I've seen it's normal to include a 500mA fuse in series with the 5v
2) Do you need to do anything with D+ and D- if you aren't using them? Will the caps affect the signals?
There is an open source (I think) generic logic analyser software called SUMP, I haven't really looked into it and it's designed for a specific LA but I've seen other devices using it around the traps so maybe if you conform to their protocol you would have a free GUI.
http://www.sump.org/projects/analyzer/client/
Last edited:
graynomad,
thanks for your comments - that's what I call constructive criticism!
fuse - I may look into this. Especially since I consider routing out the 5V supply to be used for a breadboard or similar. I recall there are fuses that don't simply blow but instead only heat up and become very resistive - but can't recall how they are called. If I can find such a thing for <$1 I'll put it in.
Unconnected D+ and D- - you actually found a fatal mistake in my circuit. That's what happens when your rip up and redo half your schematic (twice!) after midnight The data pins absolutely totally do get used - I need to connect them to the microcontroller now!
Open-source logic analyzer software - thanks for the link, I'll check it out! (If anybody knows other such pieces of software please let me know). If that cuts down on my development time I can get the scope out the door earlier!
thanks for your comments - that's what I call constructive criticism!
fuse - I may look into this. Especially since I consider routing out the 5V supply to be used for a breadboard or similar. I recall there are fuses that don't simply blow but instead only heat up and become very resistive - but can't recall how they are called. If I can find such a thing for <$1 I'll put it in.
Unconnected D+ and D- - you actually found a fatal mistake in my circuit. That's what happens when your rip up and redo half your schematic (twice!) after midnight
The data pins absolutely totally do get used - I need to connect them to the microcontroller now!Open-source logic analyzer software - thanks for the link, I'll check it out! (If anybody knows other such pieces of software please let me know). If that cuts down on my development time I can get the scope out the door earlier!
Ok, found them on Digikey. They have through-hole ones for ~$0.40. What I don't like is all the TH ones have very slow response time (2+ secs) - in that time the computer has already shut off its USB port due to overcurrent. Second, they have resistances of 2 - 10 Ohms - this means a significant voltage drop (depends on how much current my circuit draws - which I don't know yet; my guess is ~50-100mA). The circuit does not have a regulator, so this will impact level accuracy. I have to see if I can measure and calibrate out the supply voltage in any case (I think the 14K50 has an internal bandgap reference).
The Littlefuse 1206L050 SMD ones I was looking at have better specs, a 500mA version trips in 0.1sec (but that's at 8A, no figures for less so maybe 4A takes a day or two ), resistance 0.15-0.7 Ohms which is pretty good.
I think all PC have internal protection for the USB, but you can never have too much eh?
), resistance 0.15-0.7 Ohms which is pretty good.I think all PC have internal protection for the USB, but you can never have too much eh?
My 2 cents worth on adding polyfuse protection.
It seems to me that you are discussing adding a polyfuse primarily because you are considering routing out the 5V supply to be used for a breadboard. In this case the polyfuse really only needs to protect the USB port from overcurrent caused by something done on the breadboard and so it could be positioned in the 5V line at a point after the scope takes it's 5V supply. Here the polyswitch resistance won't effect the level accuracy of the scope and, perhaps more importantly, the scope supply voltage won't vary as dramatically depending on how much current the breadboard is drawing. For people who build the scope without the breadboard header the polyfuse would not be required.
Checkout the USB section in this document about using polyswitches off the Bourns site:
http://www.bourns.com/multifuse/docs/bourns_multifuse_white_paper.pdf
It makes the comment "The short circuit current as defined by the USB specification is 5 A. After 60 seconds of operation a maximal operating current of 5 A is allowed for a USB port.".
Based on this, the overcurrent you design for doesn't have to be the 500mA max for a USB port (50-100mA for the scope leaving ~400mA for the breadboard), it could be anything up to 5A. This gives you the room to trade off between the holding/trip currents and the min/max resistances when choosing a polyfuse and, hopefully, will allow you to choose a very common, and cheap, part. Keeping under the USB spec 5A max also means it doesn't matter that the polyfuse may take many seconds to respond.
To meet that max 5A overcurrent we will need to have 1R somewhere in the 5V line to the breadboard header. If we are going to power breadboard circuits then we know that the supply is going to be short-circuited at some time.and so we might want to be more conservative.
Using the figures for the Bourns Multifuse MF-R Series: http://www.bourns.com/data/global/pdfs/mfr.pdf
Part I hold I trip Rmin Rmax Min Qty Price on au.element14.com
MF-R020 0.20 0.40 1.50 2.84 5+@AU$0.53
MF-R025 0.25 0.50 1.00 1.95 5+@AU$0.53
MF-R030 0.30 0.60 0.76 1.36 5+@AU$0.558
MF-R040 0.40 0.80 0.52 0.86 5+@AU$0.558
- The 0.4A polyfuse has a min resistance that is too low so we would need to add an additional 0.5-1.5R resistor in-line. With an 0R5 resistor the short-circuit current would be 5A so we would need a 12.5W resistor!. A 1R5 resistor would need to be 10W.
- The 0.25A polyfuse has a min resistance that would just meet the 5A max for the overcurrent and would drop the breadboard supply by up to 0.2V@100mA and 0.5V@250mA at Rmax.
- The 0.20A polyfuse has a min resistance that is a little more conservative, giving a max overcurrent of 3.33A at Rmin with a drop in the breadboard supply of 0.3V@100mA and 0.6V@250mA at Rmax
The figures for the Littelfuse 60R Series have slightly higher resistances for the same hold currents.: http://www.littelfuse.com/data/en/Data_Sheets/Littelfuse_PTC_60R.pdf
Part I hold I trip Rmin Rmax Min Qty Price on au.element14.com
60R020 0.20 0.40 1.83 4.40 5+@AU$0.84
60R025 0.25 0.50 1.25 3.00 5+@AU$0.84
60R030 0.30 0.60 0.88 2.10 5+@AU$0.84
60R040 0.40 0.80 0.55 1.29 5+@AU$0.84
- The 0.20A polyfuse would have a max overcurrent of 2.7A at Rmin with a drop in the breadboard supply of up to 0.44V@100mA and 0.88V@200mA at Rmax
- The 0.25A polyfuse would have a max overcurrent of 4A at Rmin with a drop in the breadboard supply of up to 0.3V@100mA and 0.75V@250mA at Rmax
It looks that the trade-off here will be a three way one between how many extra components need to be paid for, how conservative do we want to make the max overcurrent and how much of the remaining 400mA available from the USB socket do we want to provide to the breadboard. To paraphase, "the lowest cost and fewest extra parts, having a conservative max overcurrent or providing the max available current to the breadboard, pick any two".
If you do want to put the polyfuse upstream of the scope then there are products produced specifically for USB products which have lower resistances. The Littelfuse Series RLD-USB is one such. The part with the lowest hold current is 0.75A, trip current 1.3A, Rmin=0.1 Rmax=0.23. I don't know what the availability or cost of these are compared to the standard polyfuse parts.
It seems to me that you are discussing adding a polyfuse primarily because you are considering routing out the 5V supply to be used for a breadboard. In this case the polyfuse really only needs to protect the USB port from overcurrent caused by something done on the breadboard and so it could be positioned in the 5V line at a point after the scope takes it's 5V supply. Here the polyswitch resistance won't effect the level accuracy of the scope and, perhaps more importantly, the scope supply voltage won't vary as dramatically depending on how much current the breadboard is drawing. For people who build the scope without the breadboard header the polyfuse would not be required.
Checkout the USB section in this document about using polyswitches off the Bourns site:
http://www.bourns.com/multifuse/docs/bourns_multifuse_white_paper.pdf
It makes the comment "The short circuit current as defined by the USB specification is 5 A. After 60 seconds of operation a maximal operating current of 5 A is allowed for a USB port.".
Based on this, the overcurrent you design for doesn't have to be the 500mA max for a USB port (50-100mA for the scope leaving ~400mA for the breadboard), it could be anything up to 5A. This gives you the room to trade off between the holding/trip currents and the min/max resistances when choosing a polyfuse and, hopefully, will allow you to choose a very common, and cheap, part. Keeping under the USB spec 5A max also means it doesn't matter that the polyfuse may take many seconds to respond.
To meet that max 5A overcurrent we will need to have 1R somewhere in the 5V line to the breadboard header. If we are going to power breadboard circuits then we know that the supply is going to be short-circuited at some time.and so we might want to be more conservative.
Using the figures for the Bourns Multifuse MF-R Series: http://www.bourns.com/data/global/pdfs/mfr.pdf
Part I hold I trip Rmin Rmax Min Qty Price on au.element14.com
MF-R020 0.20 0.40 1.50 2.84 5+@AU$0.53
MF-R025 0.25 0.50 1.00 1.95 5+@AU$0.53
MF-R030 0.30 0.60 0.76 1.36 5+@AU$0.558
MF-R040 0.40 0.80 0.52 0.86 5+@AU$0.558
- The 0.4A polyfuse has a min resistance that is too low so we would need to add an additional 0.5-1.5R resistor in-line. With an 0R5 resistor the short-circuit current would be 5A so we would need a 12.5W resistor!. A 1R5 resistor would need to be 10W.
- The 0.25A polyfuse has a min resistance that would just meet the 5A max for the overcurrent and would drop the breadboard supply by up to 0.2V@100mA and 0.5V@250mA at Rmax.
- The 0.20A polyfuse has a min resistance that is a little more conservative, giving a max overcurrent of 3.33A at Rmin with a drop in the breadboard supply of 0.3V@100mA and 0.6V@250mA at Rmax
The figures for the Littelfuse 60R Series have slightly higher resistances for the same hold currents.: http://www.littelfuse.com/data/en/Data_Sheets/Littelfuse_PTC_60R.pdf
Part I hold I trip Rmin Rmax Min Qty Price on au.element14.com
60R020 0.20 0.40 1.83 4.40 5+@AU$0.84
60R025 0.25 0.50 1.25 3.00 5+@AU$0.84
60R030 0.30 0.60 0.88 2.10 5+@AU$0.84
60R040 0.40 0.80 0.55 1.29 5+@AU$0.84
- The 0.20A polyfuse would have a max overcurrent of 2.7A at Rmin with a drop in the breadboard supply of up to 0.44V@100mA and 0.88V@200mA at Rmax
- The 0.25A polyfuse would have a max overcurrent of 4A at Rmin with a drop in the breadboard supply of up to 0.3V@100mA and 0.75V@250mA at Rmax
It looks that the trade-off here will be a three way one between how many extra components need to be paid for, how conservative do we want to make the max overcurrent and how much of the remaining 400mA available from the USB socket do we want to provide to the breadboard. To paraphase, "the lowest cost and fewest extra parts, having a conservative max overcurrent or providing the max available current to the breadboard, pick any two".
If you do want to put the polyfuse upstream of the scope then there are products produced specifically for USB products which have lower resistances. The Littelfuse Series RLD-USB is one such. The part with the lowest hold current is 0.75A, trip current 1.3A, Rmin=0.1 Rmax=0.23. I don't know what the availability or cost of these are compared to the standard polyfuse parts.
Last edited:
Sump is used by DangerousPrototypes.com for a couple of their designs. The FPGA-based Openbench Logic Sniffer logic analyzer, which is not really relevant here except that their comments on the Openbench Logic Sniffer page are a testimonial for Sump. More interestingly for this thread, they also use Sump for a logic analyzer mode of their PIC-based Bus Pirate http://dangerousprototypes.com/2009/11/03/bus-pirate-logic-analyzer-mode/.
The Bus Pirate page make these comments about adding support for Sump "SUMP follows a simple protocol. We’ve only implemented the minimum command set: reset, run, ID, speed (divider), samples, and trigger. Other commands are received, but the contents are ignored. Trigger direction, pre-sampling, and other advanced features could be handled with an update."
The Bus Pirate code provides a working example of a microcontroller providing data to Sump which might help further illuminate the Sump protocol documentation: http://www.sump.org/projects/analyzer/protocol/
The Bus Pirate page make these comments about adding support for Sump "SUMP follows a simple protocol. We’ve only implemented the minimum command set: reset, run, ID, speed (divider), samples, and trigger. Other commands are received, but the contents are ignored. Trigger direction, pre-sampling, and other advanced features could be handled with an update."
The Bus Pirate code provides a working example of a microcontroller providing data to Sump which might help further illuminate the Sump protocol documentation: http://www.sump.org/projects/analyzer/protocol/
I have an assortment of SMT Polyfuses here somewhere, and to the best of my recollection they are large enough for easy soldering on pcb's. After all, they only have two pins, not a hundred.
Check them out re size before forgetting about them. Some surface mnt parts are easier to mount than through hole components. These may be among them.
Check them out re size before forgetting about them. Some surface mnt parts are easier to mount than through hole components. These may be among them.
Hi all,
thanks for your thoughts and comments - very helpful. I had a short look at the Sump pages and at least at first glance found them a "bit" overwhelming - "...and where do a actually get the program without setting up a whole tool chain for recompiling???" Probably need to dig a bit more.
Anyway, got the Microchip development kit in the mail and so hopefully can continue with firmware development for the PIC18F14K50 during the coming week (after I have done my taxes :-(
Wolfgang
thanks for your thoughts and comments - very helpful. I had a short look at the Sump pages and at least at first glance found them a "bit" overwhelming - "...and where do a actually get the program without setting up a whole tool chain for recompiling???"
Probably need to dig a bit more. Anyway, got the Microchip development kit in the mail and so hopefully can continue with firmware development for the PIC18F14K50 during the coming week (after I have done my taxes :-(
Wolfgang