Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Diverted on May 29, 2018, 05:56:32 pm
-
Hi all,
A friend gave me a massive 135w Super Reverb 135-watt model with the ultralinear output transformer.
He told me one of the tubes is red-plating. I haven't done anything with the amp yet other than a quick test of the bias voltage (it's at around -55 vs 62 spec). Going slow with this because I'm not familiar with this circuit at all.
Anyway, I am wondering if anyone here has biasing suggestions for this amp. It doesn't have a bias pot but instead a balancing pot that adjusts between the two pairs of 6L6s. He doesn't mind modding the amp as it's been played with and worked on before. I'll probably install a bias pot instead of the non-adjustable fixed bias it has (recommend or no?). Also will probably replace the components on the bias board (cap, resistors and diode). They're all original.
-
If this model has 1 ohm resistors in series with the tube cathodes, you can check the voltage across those. It directly corresponds to the current through the tubes. This is what you should be worried about, rather than bias voltage (which is just a means to adjust the current). If it does not, that's the first mod I would consider because you need this to set the bias correctly, or even understand if it's in the right ballpark.
If a single tube is red-plating the most likely causes are that tube's grid resistor (high resistance or no connection to the bias supply), or a bad tube. If the bias supply produced the wrong voltage it would affect all tubes equally, though it's not a bad idea to check and overhaul it.
You need to check plate voltage to determine the target current. If it's around 500v you'd aim around 42ma per 6L6GC or 82ma per pair of tubes, for example. You want it to be balanced between the two sides, and if you can get in the right ballpark and balanced without making major changes to the bias supply, you could probably just replace bad or suspect components and not bother adding another bias control.
-
My 2 cents.
1. Always suspect tubes 1st. But we don't want to fry them in a known redplating situation. Test them in a tube tester or another amp if you can.
2. Pull the power tubes. Test for proper bias voltage on ea G1 lug. Per the tube chart @ 450 plate volts (loaded) that's -37VDC. That should be about right in your amp. If bias voltage is low, work backwards to the bias balance pot input lug. Per the schematic you should read -58VDC there. If the bias voltage reads bad, then rebuild the stock bias circuit.
3. I would mod the bias circuit only after I know that the stock circuit is working properly.
-
Get the facts straight first! Subject says Twin Reverb. Your first sentence says Super Reverb. Big difference. Which is it?
Next, fix the amp just as it is. Consider improvements only when the amp is working properly.
-
Get the facts straight first! Subject says Twin Reverb. Your first sentence says Super Reverb. Big difference. Which is it?
Next, fix the amp just as it is. Consider improvements only when the amp is working properly.
Sorry, a typo on my part. It's a Twin Reverb 135 ultralinear. Schematic here:
http://schematicheaven.net/fenderamps/twin_reverb_sf_135_schem.pdf
-
I was just messing with you. But the advice is the same. Fix the amp just as it is. Consider improvements only when the amp is working properly. That bias balance circuit is not likely causing a tube to redplate. There's a good chance you simply have a bad tube. Measure the voltage on pin 5 of all four 6L6s. If they are all close then the bias circuit is OK.
-
I couldn't agree more that there's been a general collapse of straight facts lately. I think we should all strive for improvement. (I'll wear a white hat for this post :icon_biggrin: ) It's always best to post the actual schematic.
Adding to my earlier reply: if you trace the bias voltage back another step, there should be -62VDC at the junction of the bias diode and bias cap. You don't have to trace backwards if there's good bias voltage on the G1 lugs.
Also this bias balance pot is weird tapped affair. If the pot is bad I think Doug carries one. If you mod the bias circuit, you'll need to account for blending in the tremolo voltage.
BTW: trem voltage swing can cause redplating. To test the bias circuit you may need to temporarily disconnect the trem circuit from the bias balance pot. It's possible that the trem circuit is the culprit.
-
The closest the tremolo comes to the bias circuit is through a 2.2M resistor.
-
Tremolo has nothing to do with bias! It's an opto/LDR circuit that shunts the guitar signal for the VIB channel only, prior to the channel mixing resistors, just like any AB763 circuit. The trem circuit does not connect to the bias pot.
-
I was just messing with you. But the advice is the same. Fix the amp just as it is. Consider improvements only when the amp is working properly. That bias balance circuit is not likely causing a tube to redplate. There's a good chance you simply have a bad tube. Measure the voltage on pin 5 of all four 6L6s. If they are all close then the bias circuit is OK.
Thanks. I'm going to start checking voltages tomorrow. looking at a PPP 6L6 with ultralinear, a balance pot and non-adjustable fixed bias is a new area for me, and just want to understand the most effective way to get at the issue. So thanks!
Will post some voltages tomorrow.
-
I bought a Bassman 135 when they came out in the '70's and I had a Twin Reverb 135 for awhile. They both had the same bias situation. The 1.2K resistor and the diode are on the little rectifier board. The filter capacitors and the 2.2K resistor are in the doghouse. From there a wire goes to the center tap on the balance potentiometer and a wire goes from there to the 2.2M resistor on the main board that biases the LFO way into cutoff. The 33K resistor is mounted right there at the pot. Wires come off of each end of the potentiometer and go to the 47K and 68K resistors at the end of the main board.
-
Tremolo has nothing to do with bias! It's an opto/LDR circuit that shunts the guitar signal for the VIB channel only, prior to the channel mixing resistors, just like any AB763 circuit. The trem circuit does not connect to the bias pot.
To me it looks like it enters a 4th tap on the bias pot in this amp as 2deaf says.
-
Tremolo has nothing to do with bias! It's an opto/LDR circuit that shunts the guitar signal for the VIB channel only, prior to the channel mixing resistors, just like any AB763 circuit. The trem circuit does not connect to the bias pot.
To me it looks like it enters a 4th tap on the bias pot in this amp as 2deaf says.
That's the bias voltage from the diode that connects to the tap on the bias pot. There's also an 80µF cap on that same line. No way could the trem signal survive on that line. The trem signal is on the top side of the INT pot at point B. If you look around on the schematic you will see that point B is connected to the junction of the two 220K channel mixing resistors.
-
Good point, but there still seems to be an unusual connection of opto trem & bias at the bias pot.
-
Tested out the tubes. Three of the four are weak but test above the minimum on my Hickok TV-10D/U, and one has a plate to screen short.
Will swap out a good new pair and try some voltages from here.
-
Good to know! But do test bias voltage at the socket lugs w/o tubes to make sure the amp didn't blow the tube.
-
Good point, but there still seems to be an unusual connection of opto trem & bias at the bias pot.
No there's not.
I circled all the components of the bias supply. Start at the bias diode and follow the red path to the bias balance pot. Very straight forward. Next, follow the green path to V5-2. The -58V bias ensures that the trem osc is biased into hard cutoff. Why? How does the oscillator ever work like this?
The above is the condition when the VIB footswitch is open or unplugged. This is the Trem OFF state. Now turn the Trem on by closing the VIB footswitch. This puts a ground on the left side of that 2.2M resistor, killing that -58V that was felt on V5-2. Now the oscillator can run as normal. The oscillator would work even without the -58V applied thru that 2.2M. But the oscillator would start up slowly (and get worse as the tube ages) and who wants that? When I kick a footswitch I expect something to happen immediately, not a couple seconds later. Holding V5 in deep cutoff and then suddenly slamming it into conduction by killing that -58V on the grid is like a quick "kick start". This jolt causes the oscillator to start immediately instead of slowly coming up.
That's the basic operation of any AB763 style opto trem circuit. This is no different. The whole purpose of applying the negative bias voltage to V5 grid is to ensure a fast starting oscillator, even as the oscillator tube ages.
Now consider this...
The footswitch MUST BE CLOSED for the oscillator to run. When the switch is closed, there is a ground at the left end of that 2.2M resistor. Now do you understand why the trem signal can't possibly get thru that 2.2M to the tap on the bias pot?
-
Good to know! But do test bias voltage at the socket lugs w/o tubes to make sure the amp didn't blow the tube.
Hoping it didn't, but definitely going to check.
Also, the bias diode is original and seems pretty crusty. Swap that out for a 1N4007?
Also, what would a consensus be on replacing the diode bridge on the HV secondary? It's original and the amp was putting out the right voltage from what I could see. But as a matter of course do you all think it hsould be replaced?
Thanks!
-
I would not replace either of those. The only time I replace a ss diode is when it has actually failed. Denatured alcohol and a small brush will probably shine up any crusty looking diodes if you want them to look pretty. I bet a crusty looking diode doesn't look out of place in that chassis. :icon_biggrin:
-
I would not replace either of those. The only time I replace a ss diode is when it has actually failed. Denatured alcohol and a small brush will probably shine up any crusty looking diodes if you want them to look pretty. I bet a crusty looking diode doesn't look out of place in that chassis. :icon_biggrin:
Got that right. Super crusty all around.
This amp has spent the last 30 years playing reggae and drinking music at seaside bars in a little island off the coast of Rhode Island. Rust abounds ... on everything metal out there.
-
The tube balance pot. has been converted to a bias adjustment pot. and the grid leak resistors re-configured to the standard Fender arrangement.
-
Sometimes people see the 33K that is attached to the pot. wiper connecting to the 100k on the Speed pot., but fail to notice that the junction of the two resistors also connects to ground. This schematic is very poorly drawn if the criteria for quality is how clearly it conveys the nature of the circuit.
-
With the power tubes pulled, I'm getting about -55VDC on the first grid (pin 5). Plates unloaded are at about 525VDC.
And 2deaf, you're correct, thanks! (I think?). The matching balance pot definitely does not follow the circuit. The wiper goes to the junction of two 220K resistors. The other end of those resistors (one for each pair) goes to the two grids per pair of output tubes, via the 1500K grid resistors.
What next? Thanks!
-
Good point, but there still seems to be an unusual connection of opto trem & bias at the bias pot.
No there's not.
I circled all the components of the bias supply. Start at the bias diode and follow the red path to the bias balance pot. Very straight forward. Next, follow the green path to V5-2. The -58V bias ensures that the trem osc is biased into hard cutoff. Why? How does the oscillator ever work like this?
The above is the condition when the VIB footswitch is open or unplugged. This is the Trem OFF state. Now turn the Trem on by closing the VIB footswitch. This puts a ground on the left side of that 2.2M resistor, killing that -58V that was felt on V5-2. Now the oscillator can run as normal. The oscillator would work even without the -58V applied thru that 2.2M. But the oscillator would start up slowly (and get worse as the tube ages) and who wants that? When I kick a footswitch I expect something to happen immediately, not a couple seconds later. Holding V5 in deep cutoff and then suddenly slamming it into conduction by killing that -58V on the grid is like a quick "kick start". This jolt causes the oscillator to start immediately instead of slowly coming up.
That's the basic operation of any AB763 style opto trem circuit. This is no different. The whole purpose of applying the negative bias voltage to V5 grid is to ensure a fast starting oscillator, even as the oscillator tube ages.
Now consider this...
The footswitch MUST BE CLOSED for the oscillator to run. When the switch is closed, there is a ground at the left end of that 2.2M resistor. Now do you understand why the trem signal can't possibly get thru that 2.2M to the tap on the bias pot?
Great explanation! Thanks. But it doesn't address the the left side of the bias pot which is the source of confusion. After the umpteenth time of looking at this schematic, I finally see that the 33K resistor on the left side of the bias pot goes to ground!!!, along with parts of the trem circuit. Seconding 2deaf, this is a poor & misleading way to draw a complex schematic. Such is life.
More importantly, sluckey is right that this trem circuit is isolated from the bias supply. And that Diverted, has solved or is very close to, solving the redplating issue.
-
With the power tubes pulled, I'm getting about -55VDC on the first grid (pin 5). Plates unloaded are at about 525VDC.
And 2deaf, you're correct, thanks! (I think?). The matching balance pot definitely does not follow the circuit. The wiper goes to the junction of two 220K resistors. The other end of those resistors (one for each pair) goes to the two grid per pair of output tubes, via the 1500K grid resistors.
What next? Thanks!
Not clear. You need to check all 4 grid pins for similar bias voltage as sluckey said before.
Also -55V seems high which would mean cold bias. When your sure of about -55 volts on all 4 pins, you can insert good tubes. Then check plate dissipation. Post if you need help with that.
-
Yes, all four first grids (pin 5) are showing about -55vdc.
As for checking dissipation, is the procedure different than a standard PP circuit as opposed to this PPP?
In other words, since there are 1-ohm resistors tied from the four output cathodes to ground I assume I can just read the cathode MV.But what kind of number should I expect to get in an amp with four output tubes?
Now to find a new pair of 6L6Gcs that match close to the other pair, which are marginal but still good. Will report back soon!
Thanks.
-
In other words, since there are 1-ohm resistors tied from the four output cathodes to ground I assume I can just read the cathode MV.But what kind of number should I expect to get in an amp with four output tubes?
***IF*** the measured plate voltage on the 6L6s is about 500V, I would expect to measure about 35mV to 45mV on each of the 1Ω resistors.
500V and 40mV will calculate to 20W idle dissipation. Since a 6L6 tube should have 30W max, 20W would represent 67%.
So, measure plate voltage. Expect to see about 475V to 500V.
Measure voltage on the 1Ω resistor. Expect to see less than 50mV.
-
I'd still recommend checking the powertube grid resistors-- all 1.5k grid stoppers and the 47k and 68k leak resistors (to bias supply). With no tubes installed one of those could be way out-of-spec (resistance too high) and voltage would still pass as the only "load" is your voltmeter. Hook up a pair of tubes with grid leak behavior (or positive grid current, if you're unlucky) and bias won't be steady if that resistance is way too high.
-
The 1.5k grid resistors check out fine, within five or 10% of spec for all of them. As for the 47K and 68K resistors those have been replaced with 220Ks and the junction of them goes to the wiper of the modified tube matching pot (bias pot). In the original circuit the 47K went to one end of the pot and the 68 to the other.
I assume the matching resistors that were added were put in place of any adjustability with the matching balance pot gone?
-
As 2deaf said earlier, this circuit has been converted to the more popular bias adjust circuit as found in the AB763 Twin Reverb amp.
-
OK, got voltages. Amp could definitely need some better, matched output tubes but:
B+ 506
AC: 122
Bias voltage: -57.2
Tube 1
Plate: 518
Current: 27.4mv
Dissipation: 14.2w (47%)
Tube 2
Plate: 518
Current: 31.6mv
Dissipation: 16.4w (54.6 %)
Tube 3
Plate: 520
Current: 36.7mv
Dissipation: 19w (63.4%)
Tube 4
Plate: 520
Current: 38.1mv
Dissipation: 19.7w (65.8%)
One question: Why would I be seeing a higher plate voltage than B+?
The weaker tubes (1&2) are the surviving pair from the amp. Tubes 3&4 are swapped in from my stash. Thanks!
-
220k is out-of-spec and could make it less reliable. Technically, with just 1 6L6GC that resistor should be no larger than 100k. I wouldn't worry about 100k per pair, but 220k per pair is very high.
The 47k vs 68k mismatch was more likely to address phase inverter imbalance, rather than bias.
You could put 1 weak tube and 1 strong tube on each side of the OPT. This will be (closer to) balanced across the OPT, which is the main reason we worry about balanced/matched tubes. The stronger tubes are still being run slightly harder, of course, but the amp/tone won't be affected much.
Higher plate compared to B+ must be a measuring error. Maybe there's ripple voltage at one or the other location which confuses your voltmeter. Or the readings were with the amp more/less warmed up.
-
Thanks for tip. Having never dealt w/a PPP amp I didn’t know if it was also important to match pairs of tubes on either side in addition to a match from one side of the OPT vs. the other. Good to know it’s not as important.
Also, regarding the 220ks shared as opposed to one in PP amps... . if it were your amp you would reduce those to half (100k or so) to account for that? Thanks!
-
Also, if you don't need all that power you could pull one tube fr ea pair; i.e., just use the two newer tubes, one in ea pair slot.
Anyway, you need to reduce your bias voltage. If the bias pot lacks the range needed to give proper plate diss, then a slight circuit change is needed. You might want to post a schematic of your actual bias circuit as modded, or confirm that it truly does exactly match "the more popular bias adjust circuit as found in the AB763 Twin Reverb amp." Note that small differences in the value of resistors in the bias circuit can make a big difference in bias voltage. Optimally, the bias pot at Center rotation should give you the bias voltage needed for the plate diss you want. That gives you max adjustability either way in the future.
-
Also, regarding the 220ks shared by two grids as opposed to one in PP amps... . if it were your amp you would reduce those to half (100k or so) to account for that? Thanks!
Because the recommended maximum for 6L6s is 100k, I would reduce that by half. In other words, around 50k, which is similar to the 47k-68k used originally.
If you want to keep those resistor values high for some reason, 100k is better than 220k at least.
However, due to the large 0.1uf coupling caps, even 47k doesn't reduce bass much. It will reduce signal a little, which seemed to be an intended effect of the original 47k-68k imbalance. But it should still be enough to fully drive the output. I'm not sure there's a reason not to restore those to stock value.
-
For the third and final time... THE BIAS CIRCUIT HAS BEEN CONVERTED TO THE AB763 TWIN REVERB BIAS ADJUST CIRCUIT. Those resistors are supposed to be 220K!
http://el34world.com/charts/Schematics/files/Fender/Fender_twin_reverb_ab763.pdf
-
OK thanks! I was wondering why I had never seen those caps in 100 K before, I’ve seen that pair of 220s many many times. That confused me just a bit. Staying put.
-
Leo Fender was very good at producing quality amps. Like most other legends, he didn't adhere too closely to the tube manufacturers' book value tube parameters. He made a gazillion PP 6L6 amps using 220K grid resistors. He also abused the max limits on voltages. Those amps are still alive today and are one of the most cloned amps ever. 220K, you bet I'm gonna use them, just like Leo did. Who cares that the RCA RC-30 Tube Manual says 100K max. :icon_biggrin:
-
Ok. Bias voltage is too high. Presumably, there is a 27K resistor at the "bottom" of the bias pot to ground. A smaller resistor
is may be needed. Here's a way to find the value:
* set the bias pot to mid rotation.
* temporarily replace the 27K resistor with a series combination of, say, a 20K resistor & a 10K pot wired as a variable resistor. (Don't use just a pot, so as not to risk bleeding all bias voltage to ground. )
* dial the temporary pot until you get a bias voltage that gives you the plate diss you like. You can go by the numbers, but also play a guitar through the amp to hear where you like the bias setting.
* make a permanent record of that bias voltage & plate diss for future reference.
* measure the combined resistance of the temporary pot at that rotation + the temporary fixed resistance.
* replace the original 27K resistor with one or two fixed resistors that = the new value.
-
Thanks for the suggestion. There's a 22K resistor to ground from the low side of the bias pot. I have a few questions:
By "bias voltage too high" do you mean too negative, as in its -57 at the moment as opposed to -38 or so that I should be shooting for? If that's the case, isn't the bias voltage too low (as in more negative?)
I'm also curious about this statement:
"dial the temporary pot until you get a bias voltage that gives you the plate diss you like."
I ask because I still have a lot of leeway on the bias pot to bias the tubes much hotter if I want (or colder) with that 22K resistor in place . Given that, what would be the benefit circuit-wise to swap out that resistor for a lower value, as opposed to just adjusting the bias with the pot itself? Since I can get proper bias the way it's configured now, isn't that enough? Or are there any issues caused by the bias voltage if left as is?
Thanks!
Ted
-
Your bias setting could be considered (slightly) cold. The hottest tube is often run up to 70% dissipation. jjasilli's suggestions assumed you were not able to set it hotter, but if you are able and chose not to, there's no need for modifications.
sluckey-- just because Fender built an amp like that, and the amp survived, doesn't mean it's easy on the tubes! But you're right, it's been good enough for thousands of owners for decades. 6L6s may not be as fussy... Marshalls with 220k grid resistors have quite a reputation for killing EL34s.
I might change those resistors but it's not necessary, if the owner understands the reliability will be the same as any AB763, for better or worse. Tubes wear out sooner or later either way.
-
Your bias setting could be considered (slightly) cold. The hottest tube is often run up to 70% dissipation. jjasilli's suggestions assumed you were not able to set it hotter, but if you are able and chose not to, there's no need for modifications.
sluckey-- just because Fender built an amp like that, and the amp survived, doesn't mean it's easy on the tubes! But you're right, it's been good enough for thousands of owners for decades. 6L6s may not be as fussy... Marshalls with 220k grid resistors have quite a reputation for killing EL34s.
I might change those resistors but it's not necessary, if the owner understands the reliability will be the same as any AB763, for better or worse. Tubes wear out sooner or later either way.
Thanks. My plan is to up the dissipation. I set it cold during the first fire up because the main issue my friend was having was redplating. I figured running it a little colder from where it was initially set was just a good safe thing to do while I figured out what the issue was. Once I swap these four tubes around to get more or less equal current on each side of the transformer (swapping tubes 1 and 3 should get me as close as possible with these four tubes) I'm going to bias it around 65 percent or so.
Thanks!
-
By "bias voltage too high" do you mean too negative, as in its -57 at the moment as opposed to -38 or so that I should be shooting for? If that's the case, isn't the bias voltage too low (as in more negative?)
Your bias voltage is too high. I promise you, -57v is more voltage than -38v. Think about it like this... -12V won't even tingle if you touch it but -150v will rattle your teeth.
Your bias for those 6L6s will likely end up around -50v on pin 5 when you get the idle dissipation set to 65-70%. 6L6s will probably not be happy with only -38v.
-
sluckey-- just because Fender built an amp like that, and the amp survived, doesn't mean it's easy on the tubes! But you're right, it's been good enough for thousands of owners for decades. 6L6s may not be as fussy... Marshalls with 220k grid resistors have quite a reputation for killing EL34s.
I really doubt that a 220K grid resistor is any harder on a 6L6 than a 220K. As for EL34s, they have a reputation for short life in guitar amps. I doubt you can prove that short life is due to a 220K grid resistor. The tube manuals say the max grid resistor for class a and ab is 700K! Class B is 500K!
-
By "bias voltage too high" do you mean too negative, as in its -57 at the moment as opposed to -38 or so that I should be shooting for? If that's the case, isn't the bias voltage too low (as in more negative?)
Your bias voltage is too high. I promise you, -57v is more voltage than -38v. Think about it like this... -12V won't even tingle if you touch it but -150v will rattle your teeth.
Your bias for those 6L6s will likely end up around -50v on pin 5 when you get the idle dissipation set to 65-70%. 6L6s will probably not be happy with only -38v.
Great. Thanks. I see your way of looking at it.
Looks like I should hopefully be good to go once I balance each side of the OPT and re-bias. Will report back with numbers. Many thanks.
-
I frequently qualify the magnitude of a negative bias voltage as an Absolute Value and that usually eliminates all confusion. I have been less successful in specifying the state of a standby switch.
In absolute values, is -57V the largest bias voltage the bias circuit will produce?
-
Get the facts straight first! Subject says Twin Reverb. Your first sentence says Super Reverb. Big difference. Which is it?
Next, fix the amp just as it is. Consider improvements only when the amp is working properly.
For the third and final time... THE BIAS CIRCUIT HAS BEEN CONVERTED TO THE AB763 TWIN REVERB BIAS ADJUST CIRCUIT. Those resistors are supposed to be 220K!
http://el34world.com/charts/Schematics/files/Fender/Fender_twin_reverb_ab763.pdf (http://el34world.com/charts/Schematics/files/Fender/Fender_twin_reverb_ab763.pdf)
I would like to seize upon this as a soap box moment. If a picture is worth a thousand words, then the wrong picture is worth a thousand wrong words. The source of confusion is the failure or refusal to post an accurate schematic.
It has been a pleasure to help Diverted, and I'd like to be able to give him a pass in this regard. But about half the replies in this thread are inapplicable, with members chasing through the rabbit holes of a complex schematic which schematic turns out not to match the amp in important ways, which are crucial to the bias issue in question. Also, upon popping the lid it should have been obvious that this amp was previously modded. But this info was not revealed until late in the thread. This can happen now & then. No big deal.
The larger point is that this is far from an isolated example. Numerous posts are vaguely stating that my amp is "like" this other amp; or is "based on" the attached schematic(s). These analogies suggest how an amp sounds or performs on stage, but are meaningless when it comes to the technical aspects of diagnosing & fixing an internal problem. This is a technical Forum. We should not go forward without a proper schematic. (Like a doctor would not diagnose a bone fracture without an x-ray.)
And:
I frequently qualify the magnitude of a negative bias voltage as an Absolute Value and that usually eliminates all confusion. I have been less successful in specifying the state of a standby switch.
In absolute values, is -57V the largest bias voltage the bias circuit will produce?
Ditto to 2deaf. In common parlance Negative vs Positive is ambiguous, because it can mean different things in different contexts. Within our context -- electricity or electronics -- it is not ambiguous. A larger Number is a larger voltage. The (-) or (+) signs denote phase or direction. So, -2VDC is not a smaller voltage than +2VDC.
-
Thanks for the clarification on the voltages. It makes perfect sense.
As for the schematic being inaccurate based on the amp in front of me, I'm sorry but in my defense I posted it in good faith, before I had torn open and closely looked at the amp, before I knew that the bias circuit had changed. I'm sorry the whole thing turned into a rabbit hole of inaccurate info; this was not my intent.
I'm going to go ahead and draw out as much of the power/bias supply schematic as I can, looking at the amp. My schematic drawing skills are going to suck but hopefully the diagrams I post will help clear things up.
Again, sorry. This forum and the users here are an invaluable help to me. I didn't mean to cause posters to pull their hair out.
T
-
Again, sorry. This forum and the users here are an invaluable help to me. I didn't mean to cause posters to pull their hair out.
Hey man, you're cool. Check out tyler over on TAG for an example of an Adam Henry.
From what I can see and your descriptions, I'm pretty sure I already know the bias circuit. The main reason I asked about the maximum negative voltage was to verify my hunch.
Notice, also, that the Master Volume Switch has been removed as is typical because it was a really ill-advised venture on CBS's part. The tech left the wires and capped them off as opposed to removing them, possibly so that someone could easily resurrect the switch in the future. The more crucial carbon comp. resistors have also been replaced. It appears to me as though the tech knew what he was doing and this amp is in excellent electronic working order.
I would like to see/hear about the doghouse.
-
Thanks! I don't know the max negative voltage; didn't dial it all the way over. But I will post it later.
As for the doghouse, here you go. Looks like he/she did a good job last time. Thanks!
Ted
-
That confirms that my hunch was wrong. He has replaced the 2.2K resistor with a 1.5K. I can't make out what the current-limiting resistor is on the diode board, but it is a new one.
-
I think that only one person fell in the rabbit hole. Some others may have reached in to pull him out. :wink:
-
New voltages.
I need to get some better matching tubes. Though current's pretty close on either side of the OPT, the imbalance in each of the pairs probably is forcing me to run this too cold overall, since I'm biasing for the strongest tube. Probably needs to come down a little bit overall (38 too high?) but if I come down the weaker tubes are going to be run super cold.
Anyway, the bias pot goes from about -40 to about -61vdc.
With grids seeing -58.4vdc, I'm getting:
Pair one: 30.7&36.1 for current, for a total of 66.8
Pair two: 38.4&27.1 for current, for a total of 65.5
-
I can't make out what the current-limiting resistor is on the diode board, but it is a new one.
That resistor at the bias diode is 1K.
Thanks!
-
Anyway, the bias pot goes from about -40 to about -61vdc.
That's a perfect range for 6L6s. Whoever did this conversion knew what they were doing.
So, what were the plate voltages for those last current readings?
-
Sorry, got sidetracked with work right in the middle of message. Hit "post" and forgot about plate voltage.
Anyway, plate on first pair is 523vdc. Second pair is 521.
Looks like:
53.5 percent max dissipation on tube 1.
62.9 percent max dissipation on tube 2.
67 percent max dissipation on tube 3.
47.2 percent max dissipation on tube 4.
Looks like I should be getting rid of tubes 1 and 4 and going with two a little stronger.
-
I think that only one person fell in the rabbit hole. Some others may have reached in to pull him out. :wink:
Not to beat a dead horse, but 2deaf, 92volts & I all went off in different directions. Nevertheless this has been a good thread.
-
Thanks! I don't know the max negative voltage; didn't dial it all the way over. But I will post it later.
As for the doghouse, here you go. Looks like he/she did a good job last time. Thanks!
Ted
I don't see any balancing resistors parallel with the series connected 220 uf capacitors.
-
I don't see any balancing resistors parallel with the series connected 220 uf capacitors.
They are not really needed since the PT center tap is connected to the junction of the two caps. But I "think" they are located just to the left of the PT.
-
OK, I see them now. I was looking on the dog house board. Thanks.
-
Yep, two 100Ks on a terminal strip to the left of the power transformer.
-
Leo Fender was very good at producing quality amps. Like most other legends, he didn't adhere too closely to the tube manufacturers' book value tube parameters. He made a gazillion PP 6L6 amps using 220K grid resistors. He also abused the max limits on voltages. Those amps are still alive today and are one of the most cloned amps ever. 220K, you bet I'm gonna use them, just like Leo did. Who cares that the RCA RC-30 Tube Manual says 100K max. :icon_biggrin:
Bear in mind the rather high HT on these amps; some of the brown series amps had HTs above 500V, notice how that didn't last for long, with the HTs getting reined back to compliant levels with the blackface series.
The grid leak limit is there to prevent grid current creating a significant grid voltage. Grid current, especially positive, tends to increase with gassy tubes, but also increases with overall temperature; hence if the plates are dissipating heavily, (positive) grid current can increase, and if the grid leak value is excessive, then the tube's operating point will get hotter and there's the risk that things can spiral out of control, as the hotter operating point causes yet more plate dissipation, and the tubes redplate. See http://www.mif.pg.gda.pl/homepages/frank/sheets/084/k/KT88_GEC.pdf how there's a balance between max dissipation and max grid leak, ie if the design can cap dissipation at a lower value, then the grid leak limit can be increased.
And similarly, self regulating effect of cathode bias allows higher grid leak values.
Hence most guitar amps 'get away' with 220k grid leaks, due to the nature of the application not tending to result in max dissipation for extended, continuous time periods, eg things get chance to cool down between notes etc.
But the 135W TR is not most amps, the HT as noted is rather high. Even with the slightly higher p-p impedance these have, and the g2 taps further mitigating stress, these amps seem to have been designed around the Sylvania STR387, which were a super tough 6L6 variant reputedly with enhanced voltage, dissipation and vibration capability.
The higher HT will likely increase the max plate dissipation that the 6L6 are subjected to, and thereby the grid current may well be higher than with a regular TR AB763.
So I suspect that regular / current production 6L6GC would have a hard time in the stock circuit, hence I don't think that 220k grid leaks are a good idea here. The SF circuit seems to have been done by competent engineers, as the range evolved, mitigation for both blocking distortion and tube life were put in place, and it seems a retrograde move to revise all component values to the AB763 equivalents without carefully thinking everything through.
-
@pdf64: Very informative!
Temporary thread hijack: I would like to improve tube lifetime in the attached schematic. Based on the GEC KT88 sheet I now plan to: i) with the amp in its cabinet take the tube temperature with an infrared thermometer & install a fan if needed so as not to exceed 250 deg centigrade; ii) replace R21 (off the HT secondary) with a thermistor, or a thermistor + R in series, that add up to the 10 Ohm spec of R21. I am a happy camper, thanks!
-
That sounds a good idea; in case of tube shorts, fuses for the 2 HT windings would be nice too :icon_biggrin:
It's handy to have a 125V primary available, perhaps they could see the future :think1:
-
It's handy to have a 125V primary available, perhaps they could see the future
:l2:
in case of tube shorts, fuses for the 2 HT windings would be nice too
Prescient! It already happened. Power tube went short and must have been like that for hours, before I noticed the smell from another room. PT was scalding hot and smelled real bad (but not a burnt or smokey smell). Took overnight to be cool enough to touch. Amazingly, still works fine!
My bad. Stock has 10 Ohm bias sense resistors which also serve as HT "fuses". I cleverly replaced those with 1 Ohm 3W bias sense resistors which dutifully survived the short circuit condition (maybe I should recheck their 1% 1 Ohm spec!). I then installed fuses between cathode > ground. Didn't think to fuse the screen supply. Good idea! BTW: I have 3 of these servicing the front 3 channels of my home theater system.