Bias measurements 6L6

[metallifornia]

Anyone know how or why i would get two different bias voltage readings on a 6l6 push pull quad?? For example the one pair will will read bias sweep from -59 to -48 and the other end from -59 to -30. I have measured with the tubes removed and they are then equal in measurement sweeping from -59 to -30. I tried different tubes just for S+G's but had the same effect. All of the resistors measure to their respective values. It has (2) new 220K bias supply resistors in parallel with each feeding the (2) 1.5K grid stoppers in parallel. I have attached the schematic. Any help is appreciated. The .047 coupling caps are new and not leaky and the 12AT7 is new as well. I re-tensioned the tube sockets and replaced the 470 ohm screen grid resistors.

-Brian

 

[metallifornia]

anybody? I know this is simply ohms law stuff here but it is mistifying....to give a little more insight, the amp has terrible sounding distortion that kicks in nicely and decays to crap then goes clean as the notes sustain. It sound like the symptoms are directly related to my problems. 2 of the 4 tubes or half of my sine wave is not turning on the power tubes until its driven hard which sounds like cold bias. I actuallly changed the grid stop resistors to a higher value and eliminated the cathode bypass caps in the preamp to eliminate blocking distortion from the list of possible causes.


help!


-B

 

[lcsaszar]

The tubes in each pair must be matched. Try singles first, not pairs until you find the culprit.

 

[metallifornia]

I have been using only two tubes one per side. according to my precision912 tube tester the tubes are matched. could this reading be inaccurate from my tester barring calibration?

 

[Tom Bavis]

It doesn't make sense with the circuit that you posted. So the circuit must in fact be different - there's a short or leakage or miswiring.

There's a another problem if you ask me... the grid return resistors are too high per the 6L6 data sheet - 100K max with fixed bias! And with two that should be 50K - but the driver can't DRIVE 50 K! So... you must bias this VERY cold or the tubes will probably run away. My approach would be to add 100 Ohm or more cathode resistors to each 6l6, each bypassed with 100uF or more, giving some cathode bias. These will also provide convenient points for measuring cathode current.

 

[metallifornia]

Thank you so much for the responses. This poweramp circuit is ientical to the mesa dual rectifier as all of the values are the same except that the Recto uses a 12AX7. When you say that making the feed resistors 100k and two of them would be seen by the grid of any one 6l6 as 50k??? not sure I follow you there. The at7 certainly can drive more than the higher impedance ax7 correct??

 

[Tom Bavis]

According to the data sheet, the 220K resistors should be 100K for one 6L6, 50K with two in parallel. That's with fixed bias - cathode bias can take higher values. So I'm proposing adding some cathode bias. If you lower the grid resistor value, that will eat up much of the drive from the 12AT7 driver.

You can often get away with exceeding one parameter on a data sheet IF you're not pushing the others. So exceeding plate or screen voltage rating or grid resistor limits may be OK IF you're not running near the dissipation limit. Guitar amps usually must be biased light (by hi-fi standards) to be reliable,

 

[metallifornia]

So leave the 220k's and add 100k with 100uF at each cathodes going to chassis ground.....correct? Sorry, electronics tech degree intact but I am new to tubes. They don't teach it anymore. How will adding cathode bias change my problem? what is the lowest impedance an at7 can drive?

 

[richwalters]

Seems maybe something oscillating way in the RF, which is going past unnoticed and is giving you duped figures. It need not be a saturating type of oscillation but simply resident in a small way and knocking the figures. I'm tempted to try without the input tube and see if output stage DC conditions change.

Even if equipment is out of date calibration and is working well, things won't drift so far out ...a bit like an oil change, the engine won't packup and will still run for another 100,000.

richj

 

[cerrem]

For now...leave the 220K resistors in place....I know what the data sheet says...but that is not your problem right now.... the 220K will work fine for now...
First thing you need to do when troubleshooting this type of problem is to DISCONNECT the feedback wire.....
The feedback is an "error correction" signal.... It gets applied to the tail of the phase inverter...this feedback is trying to put the amp into AC balance by pushing the Phase Inverter in opposition to what may be out of balance.... Examine the amp without feedback...you really need a scope to see what the heck is going into and out of the phase inverter as well as out of the output of the amp... For basic guitar amp work...you can just get a cheap used one of ebay.... For sig gen...either use your computer or you can use a CD player with a test tones CD they sell....

Chris

 

[Tom Bavis]

I had thought of an oscillation too, but didn't mention it...

The resistors should be 100 Ohm (not 100K) and 1W. It can help stabilize the bias, and gives you a place to measure idle current per tube. Matching on an emission tester doesn't mean much... matching cathode currents at operating conditions is what you want.

The 12AT7 driver will deliver less voltage with a 50K load... but is it enough to reach full output? Need to analyze the whole circuit.

 

[metallifornia]

Ok, so I cut the feedback wire on the tranny and voila! bias back to normal and it sounds good. Anything to do with the new Hammond 1650T output tranny. It is wired accordind to the hookup diagram from hammond and is mounted at 90 degrees from the power tranny. I wired it to the 8 and 4 ohm jacks only on my amp.

 

[cerrem]

Ok, so I cut the feedback wire on the tranny and voila! bias back to normal and it sounds good. Anything to do with the new Hammond 1650T output tranny. It is wired accordind to the hookup diagram from hammond and is mounted at 90 degrees from the power tranny. I wired it to the 8 and 4 ohm jacks only on my amp.

It "could" be possible that the feedback is out of phase, but did not cause motorboating...then you disconnected it and things went back to normal... Now while the feedback is disconnected..put some kind of signal into the amp...anything you can stick a 1/4" plug end into a CD, Radio..ect..ect then plug into the amp...keep the volum level low so you don't go into clipping...Then use your AC DVM to check the AC voltage levels across each 220K resistor... Did you make sure to have a de-coupling cap at the Y intersect of the two 220K resistors to ground???? You want to make sure that each side of the Phase Inverter is roughly the same amplitude....keep in mind that with the feedabck loop open, you will not have perfect match, since you are at the mercy of the mu and the plate resistance of each triode half of the 12AT7... I looked at the schematic and the Phase Inverter cathode resistor is not readable....what value resistor are you using there??? Something like 470 to 680 ohms??
The trick is to REVERSE the input grid wires on pin#5 to change the phaseing...make sure to have the feedback wire connected then turn the amp on with the tubes cold with no standby...this way the tube will come up slowly and yo can quickly shut down the amp if you start to hear it go before it goes into violent oscillation

Chris

 

[metallifornia]

Phase inverter cathode resistor is 470 Ohm and no there is not a cap at the Y intersection of the 220K Ohm resistors. I semi trust the engineer on that one. ((keep in mind this is a 1989 Randall RGT 100 used by the likes of Dimebag Darrell Abbott and TESLA!! Haha!! all kidding aside, this amp was once a perfectly functional beast in the '80s when it rolled off of the assembly line in Irvine, California.)) So, back to the out of phase scenario, you say that flopping the Grid wires on pin 5 of each pair of 6L6's will potentially put the signal back in-phase? Is this due to the output tranny being wired wrong or is it because of inevitable circuit oscillation due to me changing the output tranny?


-Brian

 

[cerrem]

If this was once a perfectly running amp....
Then you change the OPT...there is a chance the primary wires on the new OPT may be backwards....this puts the output of the amp 180 out of phase...not really a big deal until you apply closed loop feedabck..then there is a problem... I am not sure if this is the problem or not..so to correct this phse issue...youcan either reverse the primary transformer wires....Pin#3 to Pin#3 or you can cross swap the pin#5 from both tubes...either way gets the same job done... You are probably better just swaping pin 3...where youconnected the OPT .... Are you 100% sure that you have the primary wires correct....ie, you did not accidentally use the Center-Tap to connect to one of the output tubes??? I have seen that happen...
The way to confirm ios to swap the wires then connect the feedback wire..then turn amp on....If it oscillates violently then have your hand on the switch ready to shut it down fast obviously go back and put the wires back the way you had them... If the amp turns on nicely, then play through it and see how it sounds.... Is the new OPT of the same primary impedance???
As for engineer??? There were no real engineers building these amps...

Chris

 

[metallifornia]

Haha!! Yeah I guess you're right about that. I'll try the OT primary swap and repost my findings.

Thanks again all!!

-Brian

 

[richwalters]

If it oscillates violently then have your hand on the switch ready to shut it down fast obviously go back and put the wires back the way you had them... If the amp turns on nicely, then play through it and see how it sounds.... Is the new OPT of the same primary impedance???
As for engineer??? There were no real engineers building these amps...

Chris

There's a less painful way of amp howling; if you aren't sure how amp reacts with global nfb, temporarily connect a high value resistor, say 47K and if volume increases then one set of tranny windings (prim or sec) is wrong way round.

As for no real engineers building amps, true, many amps I've come across are close unstable i.e squeal and whistle when switched on and off.

r

 

[metallifornia]

I added 100 Ohm cathode resistors to each 6L6 then did a complete bias set up with 3 hours of run-in. I then reconnected the cathode bypass caps in the preamp stages except for the final stage (V4B). The reason is beacasuse it made the distortion harsher sounding on the dirty channel. Regardless I left it off. The amp has a little less gain in the final preamp stage because of it which is going to suit my use for this amp better anyway. I will only be using it as a clean poweramp that I will run guitar effects through. I then reconnceted the feeback wire and the sound got a little harsh and the volume dropped slightly so I removed it. It does not sound anywhere near as bad as it did when i first posted and it has all together stopped "Farting". To be honest I am SUPER pleased with the sound the way it is and the tubes are well under power dissipation limits. That being said, is it harmful to run it just like this since i am so pleased with the sound??

I am bidding on a Tektronix 922 on ebay that ends in a few hours. I will dig in further to get the crossover distortion minimized and the at7 vs 6l6 clipping further refined when i get the scope.

Thanks again you guys!!! I think I found another facet of electronics that I love........tubes!!

-Brian

 

[richwalters]

It does not sound anywhere near as bad as it did when i first posted and it has all together stopped "Farting".
-Brian

Reading between your lines I detect symptoms of instability in your amp. The "farthing" could be due to excessive interstage coupling cap values, when an output transformer is used with extended response/size and lower inductance. A symptom and influence of quality, yes.
Output transformers are the most fussy items in tube amps. A full blooded correctly working amp with 20dB global nfb inversely connected will howl with so much protest that one has a power oscillator.
Yours appears as I mentioned in an earlier post a stability or reactance oscillation which appears when global nfb is the correct way round. Without a oscilloscope one is clueless to go further. I have no industry experience of Hammond transformers, how good or bad the leakage parasitics are. To get into this further, we have to do some math.

To be careful working on an amp verging in instability, the HF ringing may also be causing the LF to "fart"when the output load i.e LS crossover is reactive. A power tube under high self RF drive usually shows getting "ruddy" hot.
I will re-examine your schematic.

richj

 

[richwalters]

Fodder worth reading. Radiotron handbook 4th edit (neg feedback) page 352 onwards. A similiar circuit is shown, except yours uses a switched R/C network in the global nfb line. I can't 'make out the wording.
However, when switched in will surely alter the phase characteristics of the amp and as Cerrem mentions circuit obviously worked for specific transformers but behaves critically with transformes from other manufacturers.

Solutions ?

richj