• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

High voltage driver for AB2 operation GU81m tubes

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
couple of questions:

The gu81m are DHT tubes with a physical ceter tap (pin 5).
What is the best place to connect pin 5? SIGNAL GND, PWR GND, CHASSIS (safety GND).
Is some resistance required or capacitor to achieve better performance?

thanks

Are you saying that the centre-tap is on the filament? If so I presume that this gives the option of series or parallel filament operation. In this case I would opt for parallel operation if using DC filament supply.

Centre-taps are notoriously 'un-centred.' For AC operation, probably the best bet is to use centre-tapped filament transformers with series operation of the filament.. The centre tap of the transformer is then taken to 0V assuming fixed bias or to the cathode bias resistor otherwise.

Paul
 
Last edited:
Thanks 7N7! Yes the filament center is a physical connection. The transformer I had made (previous posts) is 12.6v center tapped. Right now I connected all the CTs to 0v. The gu81m obviously works even if pin 5 is lifted but I think it should be grounded. Strangely enough pin 6 (g3) was originally connected to pin 5 (filament CT) via a "hook". This is how I found the original sockets.

BTW:

front panel completed! 47k log pot and 1M log pot

2012-06-25183849.jpg


powdercoated panel, connectd to chassis GND (I brushed some coating away to allow for the connection).

center connectors are high current banana plugins (each channel's connections are "vertically stacked" so imagine a vertical line in the middle and you will have ideally divided the two "halves" of the amp (right and left channel).
 
Last edited:
For G3 the choices are either to ground it or to charge the plate with small voltages. there was some interesting research by smoking-amp on the topic.

what is the benefit of connecting the suppressor to the center tap and NOT ground them both? :)

Not certain, but I have an idea that referencing g3 to the cathode might be a good idea; I have forgotten why I think this. I am trying to remember where I connected g3 in my 813 p-p amplifier - it was not tied to the anode - I think actually it went to 0V; I had arranged a degree of cathode feedback from the secondary of the output transformer.

Paul
 
I have read about this too. Infact the original sockets had the "hook" between G3 and CT.

I can't think of any reason when using DHT AC powered filaments but I am no expert :)


It would make sense to keep it slightly negative but more than that is beyond my understanding. Hopefully someone else will help us with this one
 
some initial problems testing the driver boards. Will have to reqork them with much better layout, common ground plane (to all five tubes) and CCS devices for each anode plate (10: 8 supertex dn2540n and 2x IXYS 10M45s)

IMG_0014.jpg


the front panel is lifted from the chassis to allow for additional airflow (even the sockets are "lifted" to avoid heat transfer to the chassis and additional airflow)

IMG_0015.jpg


Where there is the testing heatsink, I will place the RC filters (each voltage in the amp).
 
RC filter array ready:

IMG_0020.jpg

(sandwiched between lexan to protect the operator (me) when servicing the amp
All voltages including -12v bias will be RC filtered.

Rev 2 board almost complete:

IMG_0017.jpg


IMG_0019.jpg


Improvements:

1) single large GND plane.
2) wires for heaters are now running "under" the board and not all over the place
3) each devices will be CCS loaded with a dn2540 + lm334 cascode
4) each cascode willhave its own stabilization filter (0.1uf 630v + 10r resistor) and trimmer to set current (200r multiturn)

For now that's it. Hopefully this new board will "deliver" without any problem (fingers crossed).

I scored 28 470uf 400v nippon chemicon brand new caps for 30 euros so I managed to complete the RC filter array spending little and using fresh, unused electrolytics. I hope this will lower that residual noise :)
 
Good news and bad news....

actually excellent news and disastrous news:

The amp works.

the new configuration of 6bl7 and ccs works. I tried with a piece of music fed directly to PI stage and it yielded a very acceptable result with no fb applied

video-2012-08-05-12-23-38.mp4 video by Alexontherocks_2008 - Photobucket

The video shows a very low signal applied to the PI stage. The choice of music is a tribute to the gu81m and its engineers so no politics envolved :)

Now to the disastrous part:

The first 10 or so startups where variac controlled with a very slow ramp up. No problems. The bias on the gu81m was very low. I also saw very little screen current which is wuite normal I guess.

the 15k biasing resistors for the mosfet showed a 16v drop on idle with the mosfets producing little heat


The noise floor is excellent with virtually no hum/hiss or noise through a small medium sensitivity speaker.

Now the bad part.

Upon the 11th or so startup I tried a faster startup to simulate a true, immediate, startup. The result is visible in the following picture:


2012-08-05123729.jpg


flamed out uner lead n. 4 (CT to GND for feedback).

is this oscillation? voltage surge?

Please help. :)
 
I just had a long 1 on 1 with my muay thai kicking sack..... and I really beat the crap out of it.


Right now I am aggravated as I have possibly found the reason for what happened. Hopefully someone will be able to confirm this.

for two very different reasons both g3 leads made poor connection (reminder to on self: never ever ever leave a job unfinished in the hopes of remembering to complete it the following day). With g3 floating I built a dynatron oscillator or am I incorrect?

Could this account for the damage to the OPT?

Oscillation could also account for:

a. very very faint sound emanating from the insides of the amp. small crakling but barely audible

b. some crackling in the highs while playing music.
c. undefined mids in the same track.

I thought it was due to a very crude tuning, no opt and wrong impedance mathching (6 ohm speaker on 8 ohm leads) but now I think it is all connected to oscillation.

I did install suppressing resistors on the anodes (13 ohm + 7 turns of wire).
 
I'm thinking the insulation in the output failed. What where they rated to withstand? Did it arc to ground or from primary to secondary? To me the cracking sound was probably Coronal discharge. What dc voltage where these rated to handle?And what was the b+ voltage?

But I'm no expert and someone with more knowledge will hopefully chime in.


Nick
 
The transformer should be able to withstand 3kv. I am positive it did not arc to ground as the GND CT was not connected.

The main b+ was in the 1700v region (measured across one of 5 capacitors in series). The gu81m can handle up to 3kv.

The company which built the transformer is a respectable enterprise and has performed flawlessly thus far. I am hesitant to believe the insulation failed under "normal" "working" conditions. I think the floating g3 caused some instability but I have no experience with high power transmitter tubes.
 
I'm thinking the insulation in the output failed.

Well, that much looks obvious. The real question is why?

The transformer should be able to withstand 3kv.....The main b+ was in the 1700v region

In a P-P amp under normal conditions the voltage across the primary can reach twice the B+ voltage. Think of the OPT primary as a lever or see-saw. The center tap sits at 1700 volts. If one tube is driven to hard conduction, it pulls its end of the lever down toward ground. I don't know the particular tube, but lets assume it can pull its plate down to 200 volts (probably much less). The other tube should be at or near cutoff, and the lever shouldn't bend, so its plate will rise an equal amount from the center tap. This puts its plate at 3200 volts. This puts 3000 volts across the primary.

So in normal operation we are at or above the 3000 volt rating of the OPT. Again, I don't know the behavoir of this particular tube, but if a loose G3 connection can cause an abrupt change, or worse a cesation of plate current we could have an abnormal spike in plate voltage. Anything casuing clipping, including a sporadic oscillation can also cause a plate voltage spike.

An improper load can also cause extreme plate voltages. However a lower than normal load impedance will help you, while a higher load impedance is bad. Vreify that the test speaker is still good. Something could have blown the speaker, leaving the amp unloaded, and an unloaded P-P OPT will arc over. I have seen spikes to 3000 volts on a guitar amp that had a 400 volt power supply voltage when the load was removed at full power.

I do all new amp testing with a pair of 8 ohm 500 Watt resistors for a load. Then I parallel that with a speaker that has a 500 ohm 5 watt pot in series. This way I can crank up a 500 watt amp to full power but hear it at the 1 or 2 watt level. I got the resistors from Ebay several years ago.

Either way, the transformer doesn't look too good. Perhaps someone at the manufacturer is willing to remove the tape and tell you what arced to what. If they are not willing to help, and they just want you to buy a new transformer, you need to perform the autopsy yourself to figure out what arced to what.

I have dissected a few fried guitar amp transformers. The most common place to see a flashover is where the primary leads connect to the thinner winding wire. Often the arc will start from one lead wire junction, punching through the insulation to the winding underneath. In a very few cases this can be fixed.
 
you can pull out the OPT and test insulation resistance using a megger between the primary winding and the secondary point where the flashover occured....

i have dismantled toroids before and i noticed that there are just plastic tapes and not isulating varnish were used, so that you have highly porous traffo wherein moisture can seep in and do damage....

could also have been a corona event that BudP likes to mention a lot....

imho, you have to do something to your opt's to make them more moisture proof....
 
Thanks to all for your suggestions and observations.

The team (lem8r and I) came to the conclusion that a number of factors were possibly relevant in the damage I witnessed:

a. Possibly insufficient insulation, which, hopefully can be fixed as there is enough space on the core.

b. g3 loose causing instability

c. huge inductance of the primary Inductance measured: 941 Henry. Lem8r (who designed a 400w OPT) suggested that sequenced startup and shutdown of screen voltage could help with this one

d. umbalanced tubes.

e. The variac ramping caused spikes in all voltages which, in conjunction with the floating g3 could have caused voltage spikes in the opt. (I am actually a strong supporter of this). There is some evidence.....when turning the variac the gas rectifiers display some pulsating/glowing before settling to a stable condition. Could abrupt turn on (through NTC obviously) have avoided the damage?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.