Tube Curve Tracer design

[Synchu]

And here’s another one (in French, but google translator does the job). I have his other board (the tube tester) that works ok.
https://www.radioelec.com/en/vacuum-tub ... 7-828.html
Seems a bit older tech (using RS232 interface), however, might be useful to explore
Niki

[pompeiisneaks]

I think some of that kind of thing is 'close but not exact' to what R.G. Has in mind, His is to just put the tubes in their operating range and run the for a while and gather data over time on what they do etc. Rather than just do the typical curve traces. I'm sure that is one thing it can also do, but it would be other data collection tricks it seems he has in mind.

~Phil

[Tony Bones]

Just for reference, there is a list of other tube testers/curve tracers here: http://www.dos4ever.com/uTracer3/uTracer3_links.html

[R.G.]

Wow. There are a zillion DIY tracers!!

I looked at these a bit. I still think that a new approach to categorizing tube operation may be useful though. In my mind, curve tracing is what you do when your tools are limited to paper and CRTs. Back when I had to mess with actual curve tracers and paper plots, the things to do were always to look at the chart, pick points and calculate things like transconductance, gm, mu, etc. at that point. By taking all the raw data as points, we can make charts and such of things that would be very difficult to make appear on a CRT.

My life has freed up a bit and I can get back to looking at this.

[martin manning]

I looked at these a bit. I still think that a new approach to categorizing tube operation may be useful though. In my mind, curve tracing is what you do when your tools are limited to paper and CRTs. Back when I had to mess with actual curve tracers and paper plots, the things to do were always to look at the chart, pick points and calculate things like transconductance, gm, mu, etc. at that point. By taking all the raw data as points, we can make charts and such of things that would be very difficult to make appear on a CRT.

Of course! CRT curve tracing and working off of paper plots is horse and buggy stuff. Post-processing of digital data allows automated smoothing, averaging, and calculation of any parameter you like, limited only by your ability to display and interpret the results. Some tracers are already producing characteristic plots of Gm, for example. I’ve not seen it done yet, but one could calculate and plot anode current and Gm along a specified load line to get a good look at matching in the precise region of operation.

[martin manning]

BTW, why is this thread in PCB design?

[pompeiisneaks]

Because we're looking to design the PCB that would be used to make these things I guess? not sure. I figured it was more in this realm than general amp building of the other topic areas.

~Phil

[JazzGuitarGimp]

BTW, why is this thread in PCB design?

So that I might stumble across it, of course!

This is very interesting. So much so, that I would happily get involved and do the pcb layout work. Let me know if I can be of assistance. I think my email address can be found in my profile, and this would be the quickest way to make contact.

Cheers,
Lou

[pompeiisneaks]

No rules broken, it's more of 'how well did you obfuscate your email so you don't get spam'

I'm still waiting for some components from R.G., we have a prototype board he sent me and a Pi to connect to it, just need to know the actual component values we'll need etc.

~Phil

[ER]

[pompeiisneaks]

This is the same type it seems to me that others have pointed out. The major difference here, is that they do a one shot and done style test, and report curve traces only. The idea on this tester is to monitor all values real time as the entire test suite is run, and gather longer term persistent data over time. I.e. if you wanted to burn in tubes, and run the tests for 3 hours, you could see how they performed, get tons of raw data of all kinds, plot it out, etc.

~Phil

[Murrayatuptown]

You probably saw the uTracer 3 kit is about $315 Euros via PayPal shipped to USA from The Netherlands. Ronald (Dekker) I believe sells less-than-full-kit items like uController. I don't remember if the PCB.

UTracer 6 is in some testing phase - it attempts addressing the increasing demands of users to transmitter power tube plate voltage and current level and positive grid bias region. I think '3' goes up to 400 V and '6 beta' is going to 1 kV (?).

When I look at the cost of repairing/replacing an open meter movement in a TV-7, and butting heads with people who had their testers overhauled & calibrated for $600, uTester3 looks more than adequate to me. I think one captures data from a port and can plot from Excel. There are updated software UI's etc. There is a community around the uTracer evolution not quite open source, but Ronald puts together about 50 kits at a time, as close to non-profit as possible, with his wife.

I constantly detour on DIY my own way to microscopic levels, a big part of my brick wall on project completion.

uT3 looks pretty good to me. By the time I replaced bad sockets in my puny BK tester or made a digital display for the TV-7, I'd be in a different millennium with a laptop-connected tester.

But I still like the idea of banana jacks to a pair of DMM's to replace the TV-7xyz meter & collect data by hand. Good data or bad I don't feel as confident about.
Odd electrical specs in the original meter movement make it uncommon. It has a US Gov't p/n...probably a bunch of NIS in depot warehouses somewhere. Tried a quote request from the only shady procurement/broker site I could find that didn't ask 20 questions for sales interest first...'No Bid' reply. That could mean they didn't have a source, or they weren't interested in my puny inquiry.

[martin manning]

You probably saw the uTracer 3 kit is about $315 Euros via PayPal shipped to USA from The Netherlands. Ronald (Dekker) I believe sells less-than-full-kit items like uController. I don't remember if the PCB...

This link with my uT3 experience is above in this thread: https://ampgarage.com/forum/viewtopic.php?f=6&t=21903 It had some accuracy issues early on, but most have been corrected or worked around.

[bepone]

somebody built some tester? i have several own designs, started 10 yrs ago, based on linear mosfet regulators for every electrode, analog inputs in processors, dig. display to show measured values Ua, Ug2, Ug1, Gm, Ia, Ig2, before/after, delta. Gm measurement in 0.1 or 1V increment (small tubes ECC83 type or output like EL34)
first one based on some old Atmel microprocessor , it has all ol that, prewired sockets 7-8-9 pin for popular tubes, some switches,double triode/snge pentode measurements etc (different display and menu)..all diy

20110418_214223_U-192.jpg

[bepone]

next one was expanded version, all diy from the head from pcb to the final result,
here passing to PIC microcontroler, more stabile bias regulators, more precise 0.1V reference, valve regulated high voltage, DIY tranformer wound by myself..built in potato box i really like to use this one,
preliminary startup

1.jpg

second better pcb with all the errors corrected

plocica.png

potato box from the closest market for enclosure

3.jpg

topside

6_cr.jpg

developing tube reg. pcb:

8.jpg

double tube regulator with 6BL7 tube and mosfet followers:

10.jpg