[Troubleshooting, Solved] Kikusui PLZ-603W Electronic Load

[nidlaX]

I grabbed a used PLZ-603W off of eBay last week, and I'm in the middle of doing some general testing / cleanup / teardown. Thus far, the load seems to work OK in constant current mode up to about 300 watts or so, but I've discovered a strange set of issues and would like an expert opinion if possible.

For those who are unfamiliar, this load takes up to 3 internal load modules, each of which seems to bear around 300W. For this 600W unit, there are two modules installed, a master and a slave module.

Firstly, the master module's fan does not spin up at all. I've verified that the fan itself is working fine, and I know that the slave module's fan runs continuously during operation, so this seems wrong.

Secondly, suspecting a fault with the master load module, I completely disconnected the slave module and ran the load with only the master module. With this setup, the load still boots normally and reports no errors, and the module's fan still does not spin up. However, the load behaves strangely when operating with only this module: in constant current mode, the load will set and sense the nominally specified current, but the actual current draw is only half that value. I'm not sure whether that is by design or the symptom of a fault, so I would appreciate any insight on the matter.

Thanks in advance!

[alm]

Having never used one of these units, I am definitely not an expert. The fan not running sounds like a fault to me. I would be very careful with putting a significant load on it for more than a few seconds until that fan is running.

As for your test with one of the modules removed, I believe changing the number of modules in the field was not very common (since that would change it from a 603 to a 303 model). It could be very well that the unit needs software or jumper changes to reconfigure it for single module operation. So with it still expecting two modules to be present, having the single module drawing only half the current makes sense.

[nidlaX]

So with it still expecting two modules to be present, having the single module drawing only half the current makes sense.

Having it draw half the current seems reasonable to me as well, but the sense current reading twice the actual amount seems odd.

[edpalmer42]

I've never seen one of these either.  nidlaX, you're kinda striking out on the expert help!     But in the manual, there's a somewhat cryptic item in Appendix 3 that suggests that 'Automatic Fan Speed Control' works in all modes of operation.  There doesn't appear to be any other mention of the fan in the manual.  That suggests that there's a thermal sensor that activates the fan when necessary.  If so, then the slave module might be the one in trouble, rather than the master.

Ed

[nidlaX]

I've never seen one of these either.  nidlaX, you're kinda striking out on the expert help!     But in the manual, there's a somewhat cryptic item in Appendix 3 that suggests that 'Automatic Fan Speed Control' works in all modes of operation.  There doesn't appear to be any other mention of the fan in the manual.  That suggests that there's a thermal sensor that activates the fan when necessary.  If so, then the slave module might be the one in trouble, rather than the master.

Ed

Interesting idea, thanks Ed! I'll keep that in mind when I'm debugging.

[3db]

If the unit was configured for two modules could it be that each one has a shunt and these shunts end up in parallel.
Removing one unit would therefore double the volts measured and cause the current to be halved.
Just a thought.

3DB

[nidlaX]

Turns out the fan problem was just another case of connector gone partially disconnected for whatever reason. I really need to get in the habbit of checking the simplest possible causes before trying to dive down the rabbit hole.

If the unit was configured for two modules could it be that each one has a shunt and these shunts end up in parallel.
Removing one unit would therefore double the volts measured and cause the current to be halved.
Just a thought.

3DB

That sounds reasonable. I can check once I disassemble the unit further. Kikusui America is not very willing to provide me with any technical documents, their reason being that the unit is discontinued. On the other hand, they seem to be perfectly willing to sell me a GPIB option for around $500 for a unit that's over 20 years old...

[nidlaX]

I might have a new problem. In constant voltage mode, the load does not maintain voltage at the input according to the set value. Instead, I get a value that's approximately 20% - 50% higher depending on the magnitude of the set value. The sense values shown on the load are correct.

Could there be something wrong with the load control loop? Would the control amplifier and the voltmeter use different sense methods? Since the load works fine in constant current mode, I would suspect a potential issue with the resistive divider network, but I would also expect the displayed sense voltage to be wrong.

[alm]

I would not necessarily expect them to have the same sensing method. For starters, there is only one display unit while there are multiple load modules. I would expect each load module to have its own divider network. Also, I believe it can limit both voltage and current at the same time, so there would be two control loops driving the FETs. There could also be in issue in that area.

Another thing I would look at is the remote sensing. I did not see any way of configuring the sensing method, so there are probably low(ish) value resistors between the sensing inputs and the inputs. Not sure how this would affect the regulation but not the readout, but connecting the remote sense terminals would be an easy test.

[nidlaX]

I would love to know what the default calibration value for CV gain is on this 603W model. Another possibility is that a previous owner messed with the gain setting when one of the load module connectors was disconnected.

[nidlaX]

I've performed a calibration of the CV gain range and the load seems to perform acceptably with a reasonable calibration value. I'll let this go for now, everything else seems to be working well.