I have this power supply. For the past year or so I've been running an Electro-Harmonix 5U4GB without any problems that I know of. Actually, once I thought it arced over on me once but I really didn't catch it in time to know for sure. I decided to try out a new JJ 5U4GB in the amp and on power-up, *sometimes* it will arc over. Every once in a while I *might* catch it arc-over once during several hours of play, but for the most part, it seems stable once it warms up.
Concerned about what I can do to fix this, I started searching the internet for information on rectifier tube arcing and what I found is a bit sketchy and I wanted to try and clarify some things with anyone who knows better than I.
Some questions that have come about in my research are:
Should I be taking DC from the 5V center tap instead of the filament?
I find that most power supplies tap B+ from the filament, but a few tap from the 5V center tap from the power TX.
Some suggestions that I read are to add diodes on the HT supply from the power TX. I'd prefer to not add any SS devices to the amp. This just seems like a band-aid to another problem. They didn't add diodes on amps in the '40's! If I *did* add this mod to prevent arc-over, are the diodes in red in the schematic above right?
-OR- Do I just have a bad tube? It's hard to tell here. I read that new-production rectifier tubes are notorious for arcing over / bad construction / etc. In addition, it seems JJ tubes are the worst?!?
My gut tells me that I should reduce my input capacitance from 40uF to 20uF and possibly increase the second cap from 40uF to 80uF to make up for it, or something like that? The 5U4GB spec states 40uF, and I find a lot of amps using 5U4GB rectifiers have 40-47uF input filter caps. But, I also find 10 and 20uF caps being used on 5U4's as well.
Any comments please?
Sincerely,
rt...
All of JJ's Octal production is unreliable and suspect. Your report is the 1st that even remotely hints at trouble with EH 5U4GBs. A rectifier that has arced over really should not be trusted, but you can try the uploaded SS diode tweak. Use UF4007s, instead of 1N4007s. Less switching noise, from the outset, can only help the cause.
A NOS GE 5U4GB is quite competitive in cost with current production specimens. A GE tube will be up to standard.
In the case of directly heated rectifiers, like the 5U4, taking the "raw" B+ off a CT on the 5 VAC winding results in the lowest amount of residual hum. If you want to flip/flop between directly heated types and types with cathode sleeves, take the B+ off pin 8 on the socket.
A NOS GE 5U4GB is quite competitive in cost with current production specimens. A GE tube will be up to standard.
In the case of directly heated rectifiers, like the 5U4, taking the "raw" B+ off a CT on the 5 VAC winding results in the lowest amount of residual hum. If you want to flip/flop between directly heated types and types with cathode sleeves, take the B+ off pin 8 on the socket.
Attachments
When they start making good quality octals I'll stop. So will Eli. They make a lot of junk - but (as the old saying goes) even a blind pig finds an acorn once in a while. It's nice to hear you've had no trouble - unlike the customer who wrote me today to say that his JJ 7027 quad had multiple shorts and damaged his amp. But you are an exception, not the rule.
I call 'em as I see 'em - and I see a LOT of problem JJ tubes.
Good for you. OTOH, a lot of folks have had miserable experiences with JJ made Octal tubes, including, it seems, the OP of this thread. I think inconsistent and unreliable are accurate descriptors of JJ made Octal tubes. It's a pity, as the Jamona plant CAN make good Octal tubes. For instance, older specimens of Jamona made E34Ls were excellent.
To be fair, many of JJ's recently manufactured Noval types are anything but problematic. A good example is JJ's 12AT7, which SY is quite fond of.
First, thank you Eli for you comments and suggestions. It was my impression that where one taps the B+ correlates to directly heated vs. indirectly heated rectifiers. Thank you for clarifying that for me!
I would concur that a lot of folks seem to have problems with the JJ line of rectifiers, specifically the 5AR4, which is why I added the question, "Could I have a bad tube?"
What about dropping the input cap from 40uF to 20 or 33uF? OR, do you feel the JJ is just bad?
I would concur that a lot of folks seem to have problems with the JJ line of rectifiers, specifically the 5AR4, which is why I added the question, "Could I have a bad tube?"
What about dropping the input cap from 40uF to 20 or 33uF? OR, do you feel the JJ is just bad?
jesus ...i just bought a set of 7591 JJ's hope they will be ok for awhile until i get real ones...
10 years ago or more IMO the best sounding 5ar4 tube is sovtek even better then NOS whatever you may have but the damn tube would ark until i did the diode mod explained above it will fix it ..see no body explained this but your cap right after the rectumfinder looks like a direct short until it builds some voltage ...hence the ark over
Lawrence
10 years ago or more IMO the best sounding 5ar4 tube is sovtek even better then NOS whatever you may have but the damn tube would ark until i did the diode mod explained above it will fix it ..see no body explained this but your cap right after the rectumfinder looks like a direct short until it builds some voltage ...hence the ark over
Lawrence
The JJ is bad and in the case of the 5U4 there are so many good examples of them knocking around at hamfests all over the country for short money I can't grasp why anyone wouldn't just buy one of those tubes. 
Just me.
I've had a few KT88 explode in Citation II with fairly catastrophic results, not that they did not sound pretty good right up until the moment they did. Lots of collateral damage. I've found their smaller tubes and 300B work just fine so I don't get why they have so much trouble with octals..
Just me.I've had a few KT88 explode in Citation II with fairly catastrophic results, not that they did not sound pretty good right up until the moment they did. Lots of collateral damage. I've found their smaller tubes and 300B work just fine so I don't get why they have so much trouble with octals..
Because far too many current production vacuum rectifiers seem to have problems meeting their PIV specifications. The SS diodes provide some badly needed PIV headroom.
All I can tell you for sure is the tweak stops arcing in some current production stuff. Is the arcing occurring because the vacuum is not "hard" enough? Dunno. The SS diode tweak is an empirical fix (not original to me) that works and (IMO) that's good enough.
The problem I see with these diodes is that they rectify the AC Voltage ( you push a pulse into the Anodes of the tube, sounds like a terrible idea to me ).
The other problem I see is that you could as well throw out the tube and simply connect the diodes with the + of the first capacitor. I don´t consider the extra usage of these diodes as a fix at all.
What causes some tubes to arc is either a bad vacuum ( with gas inside ), bad mechanical design ( the distance of the electrodes is too low ) or if you use a transformer which is considered to be "too good" ( winding resistance is way too low in comparison to the older transformers these sort of tubes have ever seen ) --> higher current peaks.
This is why I mentioned that you should rather use 2x resistors in series to the Anodes to increase the winding resistance of the transformer ( according to the respective datasheet ). This is a real tweak and this can really solve a issue..
Best regards,
The other problem I see is that you could as well throw out the tube and simply connect the diodes with the + of the first capacitor. I don´t consider the extra usage of these diodes as a fix at all.
What causes some tubes to arc is either a bad vacuum ( with gas inside ), bad mechanical design ( the distance of the electrodes is too low ) or if you use a transformer which is considered to be "too good" ( winding resistance is way too low in comparison to the older transformers these sort of tubes have ever seen ) --> higher current peaks.
This is why I mentioned that you should rather use 2x resistors in series to the Anodes to increase the winding resistance of the transformer ( according to the respective datasheet ). This is a real tweak and this can really solve a issue..
Best regards,
What is the PS driving? Something is being missed here if you are using a 150ma PS and a 5H choke behind a 5U4 and still have arcing problems even with these lamented EH or JJ rectifiers. With such low current and that large of a choke behind a 40uf cap there should be no problem at all. I'm wondering if you are seeing an arc if you have a power flicker and the amp is being hot switched. And if you hot switch the amp on a regulr basis, not knowing that that can kill your rectifiers, then that may be the problem. Need more info on the rest of the amp. 20
Last edited:
I've also seen 5U4GB rectifiers arc (both NOS and current production). I suspect the problem is poor vacuum, either due to poor quality control, or sitting on the shelf for decades.
Your Si diodes will fix the problem, if you want to go that route. I've also had luck with heater-cathode type rectifiers, like the 5AR4. They have a higher PIV rating. Samples of JJ 5AR4 have worked fine for me, but I have only 4 samples, and that is not a big enough sample to say much about quality control.
If you do use vacuum rectifiers, it is a really good idea to fuse the power transformer secondary (place fuses where you've drawn the diodes). I once saw a couple of expensive transformers destroyed when NOS 5U4GBs began arcing.
Your Si diodes will fix the problem, if you want to go that route. I've also had luck with heater-cathode type rectifiers, like the 5AR4. They have a higher PIV rating. Samples of JJ 5AR4 have worked fine for me, but I have only 4 samples, and that is not a big enough sample to say much about quality control.
If you do use vacuum rectifiers, it is a really good idea to fuse the power transformer secondary (place fuses where you've drawn the diodes). I once saw a couple of expensive transformers destroyed when NOS 5U4GBs began arcing.
Hi Eli,
Very nice penmanship.
Hi Riverty,
Yup, JJ's are very suspect (= junk). Eli has everything right in his first post. Over 35 years of audio servicing backs me up on this. Both Sovtek and Electroharmonix are excellent tube brands. Stick with those.
A 5U4 is a pretty hardy tube. A hamfest is a great place to pick up used ones. Any NOS name brand (not "Original Equipment" or other cheapies) will be very good, but I would rather just buy a new one.
Hi Gorgomat,
Its a series circuit. The diodes just help with PIV issues, and the highest resistance element controls the current output. Your point about series resistance with the transformer is a good one. Series resistance helps in many ways from a conduction angle point of view, and especially to control the peak current. They also act as fuses.
20to20 has a valid point too. What is the current draw from your power supply? To properly troubleshoot your problem, we need some more detail.
For the record, I started using Sovtek as soon as Mike Mathews gave me a call to try them out. Once I tried the Electroharmonix brand, I was sold. They have been the solution for poor quality tubes. Note that Electroharmonix is a North American tube reproduction. Sovtek were Russian tubes marked with the closest North American cross. They were at one time anyway, not sure about now. That's why some folks complained about the tone. Different tube, but very good tubes. I only ever had one noisy 12AX7EH, all the power tubes have been solid.
-Chris
Very nice penmanship.
Hi Riverty,
Yup, JJ's are very suspect (= junk). Eli has everything right in his first post. Over 35 years of audio servicing backs me up on this. Both Sovtek and Electroharmonix are excellent tube brands. Stick with those.
A 5U4 is a pretty hardy tube. A hamfest is a great place to pick up used ones. Any NOS name brand (not "Original Equipment" or other cheapies) will be very good, but I would rather just buy a new one.
Hi Gorgomat,
Its a series circuit. The diodes just help with PIV issues, and the highest resistance element controls the current output. Your point about series resistance with the transformer is a good one. Series resistance helps in many ways from a conduction angle point of view, and especially to control the peak current. They also act as fuses.
20to20 has a valid point too. What is the current draw from your power supply? To properly troubleshoot your problem, we need some more detail.
For the record, I started using Sovtek as soon as Mike Mathews gave me a call to try them out. Once I tried the Electroharmonix brand, I was sold. They have been the solution for poor quality tubes. Note that Electroharmonix is a North American tube reproduction. Sovtek were Russian tubes marked with the closest North American cross. They were at one time anyway, not sure about now. That's why some folks complained about the tone. Different tube, but very good tubes. I only ever had one noisy 12AX7EH, all the power tubes have been solid.
-Chris
Guys, I'm back to I thinking i've got a bad tube and I am already sending it in for an exchange. We'll see...
Eli, I appreciate your response, and while I can see how this fix (adding 2 diodes) would work, if one adds diodes into the mix, might as well not have a tube rectifier at all at that point. Besides, I really don't want to add any SS devices in the amp.
Currently, I don't know my power TX resistance. Will check on this.
20to20, The amp is using roughly 130mA.
TavishDad, I was thinking of fusing the main B+ supply. Would that serve the same purpose?
anatech, Good to hear about your experiences with EH. Kinda backs up what I have found as well.
What next? Don't know really. I think I just have a bad tube. Everything else I have to throw at it (5U4GB tubes) just sings along fine.
rt...
Eli, I appreciate your response, and while I can see how this fix (adding 2 diodes) would work, if one adds diodes into the mix, might as well not have a tube rectifier at all at that point. Besides, I really don't want to add any SS devices in the amp.
Currently, I don't know my power TX resistance. Will check on this.
20to20, The amp is using roughly 130mA.
TavishDad, I was thinking of fusing the main B+ supply. Would that serve the same purpose?
anatech, Good to hear about your experiences with EH. Kinda backs up what I have found as well.
What next? Don't know really. I think I just have a bad tube. Everything else I have to throw at it (5U4GB tubes) just sings along fine.
rt...
The good news about hybrid tube/SS rectifier setups is that the vacuum rectifier totally dominates the behavior of the composite.
Other than stopping that blasted arcing, the "sand" is transparent. BTW, it has been claimed that the series SS diode tweak sometimes extends the service lives of well used OS vacuum rectifiers. Lacking personal experience, I can't comment about the claim. Obviously, miracles will not occur, but a few extra months of service certainly would be nice.
how does this power supply simulate in PSII ?
Maybe adding serial resistance to the transformer will help.
Maybe reducing the value of the input capacitor will help. Increasing the second filtering cap won't do much to bring up the volts though....
Maybe getting an old GE 5U4 will work wonders too. Its tough to kill an old 5U4 (but it can be done!).
btw -- nice attachement Eli.
Maybe adding serial resistance to the transformer will help.
Maybe reducing the value of the input capacitor will help. Increasing the second filtering cap won't do much to bring up the volts though....
Maybe getting an old GE 5U4 will work wonders too. Its tough to kill an old 5U4 (but it can be done!).
btw -- nice attachement Eli.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.