Max power test - JJ GZ34 vs Eico Gt Britain (Mullard?) GZ34 vs Solid State Rectifier Distortion performance in the same Dynaco ST-70
- Thread starter pac1085
- Start date
The analyzer is connected between pins 8 and 4 or 6. With Hi (positive voltage) going to the cathode pin 8, and lo going to either pin 4 or 6 depending on which one I want to test.
No, the filament is powered. That was the whole point of building that test fixture in post 174 with the 5v filament transformer in it.So you are testing the resistance between the cathode and anode, with no AC present.
5vac to pins 2 and 8 to light the filament.
positive dc voltage to pin 8 during the test (in these examples 1kv and 1.2kv)
common to pin 4 or 6 depending on which diode I want to test.
Re-read post 184.
You might also find it helpful to read up on "insulation resistance" testing in general, lots of good articles on it that will do a better job of explaining it that me.
Thanks, but if the plates are conducting, you wouldn't have giga-ohm resistance. So this measurement must be made off-cycle.
I'm just trying to figure out how JJ (one of the worst 5AR4 made) tests "better" than a real Mullard. What do those numbers really mean?
Thanks!
I'm just trying to figure out how JJ (one of the worst 5AR4 made) tests "better" than a real Mullard. What do those numbers really mean?
Thanks!
Last edited:
yes, that's how it's setup. the plate is negative in relation to the cathode. if I flip it around it conduct full current as expected.
positive voltage (vhi from 944i) goes to the cathode.
Vlo on 944i to either pin 4 or 6.
Think of it like measuring the reverse leakage of a silicon diode on a Tektronix curve tracer...
Use ohms law to convert the earlier results to current if it makes more sense.
For example:
Mullard #1 (Tek 502 scope pull)
Mullard #2 (Eico-branded, ST70 testbed)
Use ohms law to convert the earlier results to current if it makes more sense.
For example:
Mullard #1 (Tek 502 scope pull)
| Diode | Voltage | Resistance | Leakage Current |
|---|---|---|---|
| Pin 4 | 1000V | 371 MΩ | 2.70 µA |
| Pin 4 | 1200V | 92 MΩ | 13.0 µA |
| Pin 6 | 1000V | 1,354 MΩ | 0.74 µA |
| Pin 6 | 1200V | 277 MΩ | 4.33 µA |
Mullard #2 (Eico-branded, ST70 testbed)
| Diode | Voltage | Resistance | Leakage Current |
|---|---|---|---|
| Pin 4 | 1000V | 2,387 MΩ | 0.419 µA |
| Pin 4 | 1200V | 767 MΩ | 1.565 µA |
| Pin 6 | 1000V | 1,657 MΩ | 0.604 µA |
| Pin 6 | 1200V | 647 MΩ | 1.855 µA |
Lavane
Lunatic Member
There must be some fault in the internal construction of the JJ 5AR4 for them to have such a high failure rate. I had bought two at different times and suppliers. One failed after just a few hours in my VTA ST-70, and soon after the same thing in the Tubelab 300B SET amp I built. Before recently switching to a SS rectifier in my ST-70 I had been using a brown base TAD 5AR4 in the ST-70. I had a cheap Psvane 5AR4 that I had put on my 300B amp until I got another TAD or Sovtek 5AR4, but the Psvane wound up being used without issue in my 300B amp until recently when I swapped it out for the TAD.
That’s part of what intrigued me to see if there’s anything to it that can be backed up by test data. I don’t know if anyone who has tried before
Thanks! But that giga-ohm value still intrigues me. 1.) I have never seen that stat used on tubes before, and 2.) How do you measure something that large? How precise do the dropping resistors need to be?
Still learning this game. And thanks again!
Last edited:
Yeah thats easy just move the jack from cathode to either 4 or 6 whichever isn't in use.
I'm comfortable running this thing up to 1500V for the GZ34 now, the arc detection seems to work perfect. I'm not quite sure what to make of the arcs the machine has reported so far, though. BTW, the arcing is not visible since it cuts out so quickly. Seems to be safe.
New PSVANE Horizon tube arced one at 900-something volts, but on the 2nd attempt tested fine at 1.5kv (I only tested one diode so far - trying to save these for a video im working on)
The Eico Mullard when run up to 1500v measured 273/258 Mohms IR and did not arc
Both JJs arc'd. The first one 4-8 passed but 6-8 arc'd @ 1234v, but passed on the 2nd attempt and any further attempts after that (tests perfect in Roetest after the fact just in case)
The second one 4-8 passed but 6-8 arc'd @ 1337v, on the 2nd attempt it made it to 1500v but arc'd at 9.1 out of the 10 seconds into the test and aborted. It then passed the 3rd and 4th attempts fine.
Perhaps the arc detection is set too fast, it's on the fastest setting (3uS) there is a 10 and 30uS option. Or I wonder if it's some sort of impurity in the tube that gets zapped off the first time it saw that kinda voltage because I can't get them to continue to do it.
I also tested a 5r4gy for the heck of it, those are rated at 2800piv. It seems to be reflected in the results, I only tested it at 1500v but it's IR was 58/62Gohms which is a lot higher than any of the GZ34's have read.
EDIT: tested one of the JJ's from plate to plate. The machine said arc at .9 seconds into the 10 second run after reaching 1.5kv, but I ran it two more times and it registered as 10Gohms each time. I may need to adjust the sensitivity..
Last edited:
Have you tried 5U4GB rectifiers in that setup? It's almost impossible to find modern ones that can touch the reliability of the old-school US-made 5U4GBs.
My favorites are the black plate GE and Tung-Sol (USA) ones.
Thanks!
My favorites are the black plate GE and Tung-Sol (USA) ones.
Thanks!
I did try a 5u4G (no b) old Delco one, it didn't complain about being reverse biased at 1550V on either section. I do have a new JJ 5U4G I could try. Once I'm confident in the procedure/results I'll play around with more tubes. Also, the antek toroid I'm using is only rated 2A so it sagged down to 4.6VAC. I could bump the line voltage up to compensate when testing 5U4's.
Thanks! I'd be interested in US-made 5Y3GT NOS vs. new issues, Chinese- and Russian-made tubes, too.
