Yadro getting it to work

[MrFluffy]

Hi, I just soldered together my yadro finally, and installed dos on a old dead laptop.
I connected to the int4 interface fine, see the "welcome to yadro land" etc, and do a v for version number.
Then I connected it up to a cheap vernier caliper I had for the purposes.
However when I did a da or d0, the first (and only) scale reports :-
Checking device #0...
Clock : Signal
Data : Signal

And thats it. No Device: entry. If I repeat the tests about 5 or 6 times I get the error "Device: unknown type" added to the end. The int4 hasn't locked up as it will do a v or a d0 straight away, it just isn't reporting back the status from interrogating the scale.
I tried flipping data and clock lines, and its made no difference. If I pull either of them to gnd, I get a error saying they have gone L, and disconnect them and the int4 reports theyre H for high (open circuit).
With a DA, the rest of the scales report Clock and Data as H, without a noticeable delay.

Is this just the scale not compatible with yadro? New protocol? Or something amiss with the int4 in some way? It looks like the same pcb as on the yadro site but the voltage and gnd lines on the connector are in different positions too (I connected directly to the battery pads instead).

I can order some of allendale's replacement caliper heads if its a protocol error because I've read on the net that their caliper heads work, this is just a cheapie I picked up at the local diy chain.

[John Stevenson]

There are now at least two protocols for the cheap Chinese calipers, someone has moved the goalpost.

Give Lester Caine a ring.

[MrFluffy]

Thanks for the response John. It was sourced from him originally, but I've been so long building it (over 3 years now I think...) I didn't think to ring him and ask as I saw he's not doing yadro stuff any more. Ill try and get to call him on monday if I cant resolve in the meantime, I meant to follow up on your suggestion today but things got in the way and suddenly it was 6pm.

I tried in the small hours last night with a second caliper of slightly different appearance, and got "Data, unknown protocol 39 bytes in length" or something similar with the data length changing by one or two bytes, then at one point it identified it as the 24bit scale and sync'd to it, and I started the "setup mill" program which identified it, but refused to configure as it only had one axis, then when I started the "setup lathe" program it had reverted back to being a unknown scale. I tried power cycling everything including the pc and int4, and couldn't then get it back to that situation.

It may be my psu is introducing noise into the circuit at fault causing it to miss some of the bytes in the output sequence and mis identify the partially successful scale as a result, I had bypassed the bridge rectifier section and used a simple unfiltered 9vdc plug in the wall psu as I didn't have a 9vac one, and thats also powering the revmaster tacho I built at the same time. In fact typing that, it occurs to me that it may be the revmaster introducing noise into the int4 as they're mounted in the same box and powered up and I haven't added any shielding between yet. Or mains borne noise via the psu. Maybe some of the components I eliminated before the regulator were for smoothing of the rectified dc on thinking back also. It feels noise related because its intermittent too it seems.

I have a real 12v stabilized filtered psu. I will try the int4 alone on that with the revmaster unpowered to eliminate those possibilities before I ring lester so as not to waste his time. And take a look at the scales output on the oscilloscope see if I can see noise on it which might be upsetting the int4.

[macona]

Sounds like noise. Put .01uf caps across the power lines going to the scales. You may also need some on the data lines. This is where you need a o-scope.

[MrFluffy]

Last night spent rebuilding the int4 with the psu bits I omitted as I thought could see lots of noise on the lines. Then discovered the read head I was testing against was dead and giving no output. Swapped the read head for the unidentified one and the signals were clear as day but the wrong protocol.
While dealing with the business of the day, in today's post came the allandale supplies sourced caliper heads and wixey table saw strips , and they sync up and identify correctly. So potentially noise was also the culprit with the intermittent caliper.
In my haste I forgot to put the ddisp binary on the dos partition but setup.exe ran through fine.

Also, one important thing I missed, the heads have to be mounted on a scale to work! seems obvious but a few of the tests I was running against the circuit boards hanging in mid air, and wondered why they didn't work. Worth noting in case it helps someone else...

Now have to make the bracket work, inverted U shaped extrusions to protect the scales and all the other bits and bobs.