Bulk order of LSI LS7184N Quadrature Decoder IC

[popchops]

I need a few LS7184N Quadrature Decoder ICs, and they are a bit tricky to get hold of in the UK.

If I can get 10 or 20 delivered from the US, is there anybody in UK who wants one (or more) and will pay £9.00+pp for these excellent HW decoders?
https://lsicsi.com/products/ls7183n-s-ls7183n-ls7184n-s-ls7184n/

Pops

 

[erco]

I'm in the US. They appear cheap at https://www.digikey.com/en/products/detail/lsi-csi/LS7184N-S/12760890

QTY	UNIT PRICE	EXT PRICE
1	$4.89000	$4.89
25	$3.93000	$98.25
100	$3.13000	$313.00

 

[popchops]

Thanks for your reply! I know they are easily available in the US. Digikey want an additional $80 for shipping to the UK. It's a problem.

 

[oracacle]

Erco, would you be willing to ship one or more at a lower cost?
Seems like a logical solution to extortion.

 

[erco]

Sure, HTH. PM me Popchops.

 

[erco]

Is this something a Picaxe could do?

 

[popchops]

Is this something a Picaxe could do?

Yes I currently have tested on a 28X2, just this minute soldering in a 08M2 to confirm it works. It will never be as good/elegant as the HW decoder chips, but it's good enough considering that my main loop needs a minimum IR timeout of 27ms to scan for IR input. Cannot be interrupted in this period, which means my overall rate of change is limited to ~37 changes per second without missing a change. For an encoder with 32 detents, that's more than one full revolution so quite unlikely. 32 detents represents an increase in audio volume of 16 dB which is quite a lot. As you see, a picaxe SW solution capable of ~90 changes per second is quite adequate.

Nice guys at LSI did send me a couple of samples of LS7184, but it's the surface mount package (the fat-finger version LS7184N is out of stock, and has minimum quantity of 100). I've ordered a proto board with 0.05in pitch so I will try my hand at soldering in miniature.

 

[hippy]

Is this something a Picaxe could do?

I would have thought it's the ideal application for a PICAXE-08M2. It would be possible to make the CLK_UP and CLK_DN outputs toggling which could make polling for changes easier than looking for a short pulse.

#Picaxe 08M2

;     -|+V 0V|-
;     -|SI O0|- 
; A >--|X4 X1|--> CLK_UP
; B >--|I3 X2|--> CLK_DN

Symbol CLK_UP = 1
Symbol CLK_DN = 2

;     .------- A was
;     |.------ B was
;     ||.----- A now
;     |||.---- B now
;     ||||
Data %0010, ( CLK_UP )
Data %0111, ( CLK_DN )

Init:
  SetFreq M32
  Low CLK_UP
  Low CLK_DN

MainLoop:
  Do
    b0 = pins / 8
    b1 = b1 & %11 * 4 | b0
    Read b1, b2
    Toggle b2
  Loop

That is the simplest quadrature detector, just detecting A going high. Detecting A falling as well, or all transitions on A or B would simply be a case of adding extra DATA definitions.

 

[popchops]

Yes - I will use a picaxe 08M2. It's fast enough even for full quadrature detection of all edges up to 100 Hz.

Thanks Hippy.

 

[hippy]

Untested but, if you do need more responsiveness, you can use a more complicated bit mapping in 'b1' to track changes -

  Do          ;  76543210           ;        Was     Now
    b1 = b1   & %00011000 * 4       ; .---.-------.-------.-----------.
    b1 = pins & %00011000 + b1      ; | - | A | B | A | B | - | - | - |
    Read b1, b2                     ; `---^-------^-------^-----------'

That saves on a division so should be faster. Not sure if adding 'b1' four times would be quicker than a multiplication, but I suspect not.

The DATA entries then have to be adjusted to suit that -

;     .------- A was
;     |.------ B was
;     ||.----- A now
;     |||.---- B now
;     ||||
Data %0010000, ( CLK_UP ) ; A rising  - B = 0
Data %0111000, ( CLK_DN ) ; A rising  - B = 1
Data %1000000, ( CLK_DN ) ; A falling - B = 0
Data %1101000, ( CLK_UP ) ; A falling - B = 1
Data %0001000, ( CLK_DN ) ; B rising  - A = 0
Data %1011000, ( CLK_UP ) ; B rising  - A = 1
Data %0100000, ( CLK_UP ) ; B falling - A = 0
Data %1110000, ( CLK_DN ) ; B falling - A = 1