Gripmeter by Pratt Burnerd

[awemawson]

Some of you may recall last year I was trying to find a way to measure chuck gripping force of a hydraulic cnc chuck I'd installed:

http://madmodder.net/index.php/topic,10051.0.html

Well I never progressed the idea until a little while ago browsing eBay the ideal instrument appeared. Price wasn't cheap but probably reasonable for a professional bit of kit

[awemawson]

Sold as fully working. When it arrived sadly it wasn't working. The display unit showed a permanent offset of -24 kN and the load cell / transmitter unit wouldn't power up.

I contacted the seller who in fairness instantly offered a refund, but I made him an offer (1/3rd original price) for the remains to see if I might be able to fix it. It was apparent that coolant had penetrated the transmitter and it was badly corroded within. Also the internal NiCad rechargeable batteries were dead as dodos 

The NiCads were a strange size - industry standard footprint but a low profile version to squeeze into the case.

[awemawson]

A hunt around failed to find any direct replacement for the batteries - everything was too tall to fit in. Eventually I tracked down some Lithium Ion cells with flying leads that would just squeeze in.

The originals were 110 mA/H and the new ones are over ten times the capacity - charging circuit is constant current, but only 20mA so I think that it'll be ok if not left on for days

[awemawson]

I spent a little time reverse engineering the two PCB's and producing circuit diagrams so I could see what was going on.

First PCB holds the batteries and a constant current charging circuit.

Second PCB interfaces the load cell. It has a 2.45 volt precision reference (ZN404) that is buffered by an op amp and supplies a constant 2.45 v to the bridge of the load cell. The output of the load cell is fed to another op amp with a gain of 75. This amplified DC level is presented to a voltage to frequency converter (AD654), which via a PNP transistor (ZTX500) drives an Infra Red transmitter diode.

[awemawson]

Having cleaned things up and replaced the sockets (yet to replace the pins) on the PCB things started to work, I could set zero, but put a squash on the load cell and the reading went down not up!

I ended up reversing the output wires from the load cell on the PCB and now it seems to work fine. Very odd as the wire colours are in the copper etch of the board, and the load cell seems to conform to the industry standard colours 

Anyway it works, zero is achievable and trying it on the manual lathe you can not only measure the clamping force statically, but see the decrease due to centrifugal force as the chuck spins

[awemawson]

It came with adaptors for two jaw and three jaw chucks - handy as the two jaw could be used with a welding clamp during testing 

[mattinker]

You do find yourself amazing problems to solve! Looks like you've been having fun.

All the best, Matthew.

[awemawson]

Well I finally put this one to bed today - sorry forgot to take photo's. I'd found that the zeroing drifted all over the place - I'd put it down to components marginally warming after switch on, but then I found that all I had to do to change the reading, was to breathe on the PCB that holds the load cell interface. Turns out that, as the gain of 75 of the op amp is set with a 10K and a 750K resistor, any moisture from my breath on the 750 k one significantly altered it's value  Thus the reading went out the window as I was pre-setting the trim pot on the board to it's nominal centre frequency.

The cure was simple - a couple of spray coats of 'conformal coating' (which is an acrylic varnish) on the boards after which you could practically dunk them it water and get no change  (I had to mask the trim pot and board connectors so not to gum them up)

Actually there is one more minor thing to do - the wall wart charger failed and I need to get one small enough to fit in the allocated space in the case - again ebay has provided one that's 'on the way'