Pentode Testing Question

[SerpentRuss]

I recently purchased a Pico, usb-based oscilloscope; the cheapest 2-channel model. It does have an AWG (Arbitrary Wave Generator) capable of -20 to 20 volt signal generation. I have played around with this a bit and I'm really happy I purchased it. I have a scoping multimeter, one of the better Flukes, but the screen was really small and I was previously using my phone as a signal generator. It was impossible to get a consistent test signal. Being able to use a laptop is going to help my tired old eyes. I also purchased a fixed, 100x, 2KV probe so that I can look at some of the higher voltage test points if necessary.

One of the first things I'm going to do with this new toy is sort out about 30, used EL84 and 6P14P tubes I picked up on FleaBay a while back as "tested good, strong", LOL. I think they were about 50 cents each with shipping. With the scope and a dummy load, I should be able to compare the tubes against each other by keeping my test signal constant. I have a brand new J&J EL84 as a benchmark.

I'm looking forward to getting this set up. I have a VariAC feeding a dedicated HT transformer, rectifier and filter and a fixed 6.3 volt DC power supply for the heaters. I'm going to set up a simple cathode-biased socket feeding a 4K dummy load and feed the signal in directly from the wave generator at a grid leak resistor.

My question is: Is there any knowledge to be gained from rigging up a voltage drop to power the screens, or should I just test them as triode strapped?

 

[dan40]

I would think that you would definitely want to test them in pentode mode if this is how you intend to use them. That's just an educated guess though so hopefully one of the wiser members will chime in.

 

[SerpentRuss]

I guess I'll have to set up some sort of voltage divider and feed the screen from that. I'll have to fiddle around with some values. Or I could put in a choke. Still, wouldn't I need some tail load to make it effective and keep from dropping all the voltage across it and limit the current? Something like this? It doesn't show the variable DC power supply for simplicity.

 

[YellowBoots]

Still, wouldn't I need some tail load to make it effective and keep from dropping all the voltage across it and limit the current?

Nope, not required. It would have the effect of regulating the screen voltage somewhat. Whether you install something like that depends on what your end goal is here. Do you want to have the same screen voltage on every tube, or are you ok if it varies? Regulating 400V with a voltage divider is going to be very inefficient but it could be tried. What are your target voltages?

I think I would leave out the extra load resistor just accept that screen current will vary a little with each tube and that will vary the screen voltage a little. You aren’t setting up a curve tracer here, so what’s the big problem if the screen voltage isn’t regulated?

 

[YellowBoots]

To help better understand the DC voltage divider situation in the screen supply, you could imagine the screen as a 47K resistor to ground. The divider created by the 1.5K supply resistor and this imaginary resistance will set the DC voltage on the screen. This is just an estimate but will get you in the ballpark.

 

[YellowBoots]

Eek. I just saw you posted that you plan to use a 4K dummy load. That probably won’t work because it will drop a huge chunk your plate voltage. I reckon you need an OT.

 

[SerpentRuss]

Eek. I just saw you posted that you plan to use a 4K dummy load. That probably won’t work because it will drop a huge chunk your plate voltage. I reckon you need an OT.

You're right. I wasn't thinking about the load line difference between a reactive and resistive load. I decided to use a 4K Singled ended transformer loaded with an 8 ohm power resistor connected to the secondary, because I have both on hand. My LTSpice model shows that.

 

[SerpentRuss]

To help better understand the DC voltage divider situation in the screen supply, you could imagine the screen as a 47K resistor to ground. The divider created by the 1.5K supply resistor and this imaginary resistance will set the DC voltage on the screen. This is just an estimate but will get you in the ballpark.

If I don't need the tail load it will make life simpler.

I'm not trying to push the tubes hard in the test (maybe I should), I'm just trying to sort which ones are close to spent. I think I'll let them warm up for a few minutes and try a couple of different frequencies and note the the levels for each tube. I imagine a few will go in the trash.

I'm still working on a plexi-ish project right now, but I bought a transformer that will work in a parallel, single-ended EL84 amp I have planned with an EF86 pre-amp. Call it an AC8 for lack of a better name. It will be fun to pair up different "good" valves in that project and see what they sound like.

 

[printer2]

I might make something along these lines. It does not test for shorts and stuff but I will cross that bridge later. The real interesting thing is the maker of a tube tester (which is really cool how they dumb it down) using cards for each tube type. They tested a good tube in the tester to 'calibrate' the tester for each tube. In the comment section of the video he has links to all the cards.

The interesting part is the top middle text, Test Operation. With 200V on the plate and 100V on the screen the mA scale on the bottom will tell if the tube has some life or not. At about 28 mA it is still good, in the 20 mA range questionable. The white dots are places you put 'pins' in to program the tester for the proper tube. The pins make switches and the operator does not have to know how the machine operates, just put the card for your tube in, put the pins in, test your tube. The good thing is you have the operating parameters and how much current should pass for the tube to be considered good.

http://www.roehrentest.de/TUmrechungFunke_EN.pdf

A guy who rigged up his own tester and uses the card information to test with (and has all the cards scanned as well as having a text version of the cards).

Funke W19 Prüfkarten download

Nachfolgend stehen die Prüfkarten für das Funke W 19 zum download bereit. Die Prüfkarten lassen sich direkt aus der PDF-Datei in passender Grösse ausdrucken. Die Abmessungen der Karten betragen dur…

tuberocker.wordpress.com

Not a complete plan but a person with some background and good at searching should be able to figure things out.

 

[peteb]

I'm not trying to push the tubes hard in the test (maybe I should),

sounds like a great project.


what is your plan on choosing your signal level?

 

[SerpentRuss]

sounds like a great project.


what is your plan on choosing your signal level?

I'll probably adjust the cathode resistance for about 35 mA with a known good EL84 and then take output measurements for octaves 110, 220, 440, 880, and 1760 Hz. I'll have to figure out my test signal levels after I get it set up, I imagine 8 volts peak to peak might be a good starting point. Then with that benchmark, I'll cycle through the rest of the tubes.

 

[Bendyha]

One can get a bit carried away making a D.I.Y. tester. I'm surprised this one doesn't have a scope attached with a curve tracer option.

 

[SerpentRuss]

One can get a bit carried away making a D.I.Y. tester. I'm surprised this one doesn't have a scope attached with a curve tracer option.
View attachment 1086382

Ludicrous capability for us amp people. I'll start with one tube, LOL

 

[printer2]

One can get a bit carried away making a D.I.Y. tester. I'm surprised this one doesn't have a scope attached with a curve tracer option.
View attachment 1086382

I missed getting a high voltage power supply (three outputs) for tube design as well as a curve tracer.