uTracer (Micro Curve Tracer)

[award70]

Hi Martin,

Yes those are jacks for an external heater supply. I don’t really use the internal heater supply.

Alan

[martin manning]

Some updates on the uTracer 3+.

I decided to replace the current sense resistors with 6.8 Ω and the hardware limiting resistors with 1Ω to make my 3+ a "500 mA" capable device (the compliance limit calculated in the software is actually 539 mA). This is enough current to trace 6550 and KT88 at data sheet screen voltages, and gets rid of the multi-paralleled resistors used in the original "high current" 600 mA configuration mod. It's also a bit safer with respect to the possibility of doing damage to the HV switches. Turns out the uTracer3 is better at protecting the DUT than it is at protecting itself. While I was in it, I replaced the boost converter MOSFETs and the BJT transistor pairs in both switches with current part numbers. There was some odd behavior in the traces that has now cleared up. Accuracy at low current is quite good, as shown below where 100k 1% resistor traces pass very nearly through the expected I/V intersections in the 0-3 mV range, and the screen and anode channels match very closely.

This all started with Ihor and Bob (uTracer GGroup) working on low voltage high-current accuracy in uT6 and then uT3. At Bob's suggestion Ihor added two new calibration inputs to the cal file in his uTracerJS interface. One accounts for voltage drop due to fixed resistances in the anode and screen voltage circuits, and the other corrects for voltage offset in the ADC's. This improves accuracy around the knee when tracing large audio tubes and when testing vacuum rectifiers.

[martin manning]

This plot shows uncorrected and corrected error in indicated resistance tracing 100.5 ohm resistors, where the new calibration inputs substantially reduce the error. It's still quite large at very low voltages, so the best accuracy when testing vacuum rectifiers is achieved when the applied voltage is at least 15-20V. A 5AR4 should pass just over 200 mA at 16V Va-k. With this new feature vacuum rectifiers can be reliably tested directly, meaning there is no need for the extra effort involved in adding a series resistor to the test setup so the uTracer can make measurements at higher Va and Vs, and then making a separate calculation to get the actual voltage applied to the DUT.

[martin manning]

Ihor Smal's uTracerJS has gotten a few more updates: Links to documentation have been added, including a concise uT3 calibration procedure, a detailed description of the new current-dependent and ADC offset voltage corrections, and an Excel worksheet that can be used to determine the calibration inputs (the worksheet that produced the plot above).

I'm going to run exclusively with uTJS now, having replaced my old MacBook and not really wanting to buy and maintain a Windows OS on the new machine. Below are some screen captures of uTJS forms including a full trace and Quick Test results for a nice RCA black plate 6L6GC, and the calibration input form which has the documentation links.

uTJS is available here with versions for Windows, Linux, and Mac OS, and it works for both uT3 and uT6: https://boffin.nl/wp/

[markusw]

I've got some weird phenomenon going on with my µTracer 3+.

Calibration of Va and Vs is fine (measured in continuous mode between C13 and C18 and cathode).

When I run the µTracer it indicated lower voltages than what why DMM measures (the same one that was used for calibration).
The first table shows anode voltages measured versus cathode with my DMM and the voltages the µTracer indicates ( from the saved "Measurement Matrix" file) with no tube plugged in and the 10k calibration resistors not active.
The second table shows the voltages with a ECC83 plugged in (Grids -2V).
For the screen situation is the same.

Basically, there are two things I do not understand.
Why does the µTracer indicate lower voltages than what the DMM measures?
Why are the voltages measured with my DMM higher when a ECC83 is plugged in?



Any ideas?
Might these lower voltages the µTracer measures be related to the tolerances of the voltage divider resistors (R32/R33 and R18/R19)?
Might it be that the GUI subtracts around 2V-3V to take into account the voltage drop over the high voltage switch, not knowing that the µTracer is run in continuous mode?
I don't have a clue how large the voltage drop is.

Best,
Markus

[martin manning]

Hi Markus, are you running Ronald's GUI? It does not compensate for voltage drops, so you get resistive losses and reservoir capacitor voltage sag amounting to 2-3V with respect to the voltage on the reservoir capacitors. With no load that should be small. You can't measure the voltage at the anode/cathode terminals with a DMM since it appears as very short pulses. Tracing resistors is the bast way to see if the HV and current measuring circuits are working, and looking at the quick test results for a tube is the easiest way to see how much voltage sag there is. Have you tried Ihor's uT JS? It has a double acquisition mode that adjusts the set voltage up to hit the target.

[markusw]

Hi Martin,

yes, I'm running Ronald's GUI v3.12.6
I measured the voltages directly at C13 and C18 versus cathode. All the measurements were done in continuous mode, i.e. anode/screen directly connected to the reservoir capacitors C13 and C18.
I have tried uT JS and see the same.
What puzzles me is that, although the voltages I measure with my DMM during calibration are pretty on spot, the GUI indicates lower voltages already durig calibration. This is the first table.

Might it be that the GUI subtracts around 2V-3V of anode/screen voltage to take into account the voltage drop over the high voltage switch, not knowing that the µTracer is running in continuous mode? Don't know if the voltage drop might be around 2-3V though.

[markusw]

Have you tried Ihor's uT JS? It has a double acquisition mode that adjusts the set voltage up to hit the target.

The double acquisition mode helps

[martin manning]

The double acquisition mode helps

Yes double acquisition is a great feature. I makes the curves more accurate, and there is no fiddling around required to get accurate test voltages for the QuickTest.

I don't believe there is any compensation for the HV switch voltage drop. You can read about that and the additional calibrations in uTJS by clicking on the link at the bottom of the calibration form.

[bluesbreaker62]

Hi guys, I hope someone here can point me in the right direction.

I've been away from building tube amps for a while and am slowly getting into it again and wanted to test the stock of tubes that I have amassed over the years.

I have this V3 uTracer (not the V3+) that was built years ago and put away fully working. Is there any updates I should definitely consider doing before I start testing? Thanks!

[martin manning]

Was this uT3 built by you or someone else?
Are you familiar with its operation?
Does it power up?
Do you have the current GUI?

Eventually you might want to upgrade it to current uT3+ configuration, but I’d definitely get it working as it is first.

[martin manning]

...

[bluesbreaker62]

Thanks Martin! Hoped you'd chime in

It's a project I did with my brother some 15 years ago if I had to guess

I haven't dared plug it in yet, is the current GUI downward compatible with the V3 hardware?

[martin manning]

It's a very nice looking unit! As far as I know the current GUI should work. There are some selectable options to choose the version and some other settings for the current sensing resistor value.
I can recommend Ihor Smal's uTracerJS, which should also work. Since replacing my computer I have been using it exclusively. https://boffin.nl/wp/

[bluesbreaker62]

Thanks Martin! I'll give that a try and report back