Carvin Legacy 3 bias issue or design?

[EddyMunster]

I am new to this forum so hello to everyone.

I have poked around the web for an answer to this question with no luck.

I have two Carvin Legacy 3 amplifier heads and I am getting the same results from both.

I set the bias to 100ma across the bypass switch per Carvin's method on both amps.

I did a comparison and noticed the power tubes are only running about 8ma per tube. I measured the 8ma using a bias probe at each tube socket.

I'm thinking they should be reading closer to 20ma per tube.

Is anyone familiar with these particular amps or has anyone seen this with Legacy amps?

I attached a schematic and I'm not very savvy with schematics for amps.

-Eddy

[DeepBias]

8mA is not normal at all, 20mA is about 9W and that's still low for EL34, for the 60% average at ~450V plate they should be around 33mA (15W). Maybe you did the Carvin standby method wrong, but if you have bias probes don't care about the Carvin method, just bias using the probes at 33mA per tube and you're good to go (for 5881 30mA). Be careful while ajusting P11 bias pot, you can put some tape on the chassis to avoid touching it with the screwdriver, use only one hand, the other in your back. 

[astronomicum]

At 8mA per tube, that’s a total of 32mA for the power tubes. If you are biasing for a total of 100mA across the standby, that means 68mA is going somewhere other than the power tubes. That is a lot of other stuff. I would advise caution before rebiasing to 33mA per tube which will add another 100mA to the load.

[DeepBias]

At 8mA per tube, that’s a total of 32mA for the power tubes. If you are biasing for a total of 100mA across the standby, that means 68mA is going somewhere other than the power tubes. That is a lot of other stuff. I would advise caution before rebiasing to 33mA per tube which will add another 100mA to the load.

The current is going somewhere other than the power tubes, inside both amplifiers ? I'm following this thread with attention ... but it's not impossible so yes be careful that's a good advice.
 

[EddyMunster]

Thanks for all the input. I did try going over 30ma before I posted the questions and ended up frying R55 resistor. R56 looks a little cooked as well. I had over 200ma across the standby at that setting and P11 was maxed out. That is why I picked up another identical amp in order to do a comparison. I have since replaced R55 with a 2watt 4.7k I found locally. I need to be sure it's either a problem that needs fixing or is it just how these amps run by design. I can't run the bias past the 100ma for now.

[astronomicum]

So either your bias check as Carvin specifies is wrong, or the bias probes you are using are not accurate. It would be prudent to know which before proceeding. Since you burnt stuff up rebiasing to 200mA across the standby switch, which was my fear, my guess is something is amiss with you bias probe readings.

[EddyMunster]

Sorry I should have mentioned doing some verifications I did before getting hold of another amp. I have a fluke 87-3 and a Eurotubes pro-1 bias probe. Long story short they are both accurate.

[glass54]

Hi Eddy,
If it was my amp, I would fit 1R 1W 1% on all Pin 8 (Cathode) of V5A to V8A. Cheap as chips 
At least you would have some definitive DC Test Points and measure in real world terms. My thoughts 
Regards
Mirek

ps The Fluke is great

[EddyMunster]

I don't understand what you mean. Can you please clarify?

[DeepBias]

Hi Eddy,
If it was my amp, I would fit 1R 1W 1% on all Pin 8 (Cathode) of V5A to V8A. Cheap as chips 
At least you would have some definitive DC Test Points and measure in real world terms...

That's a good way to know the accuracy of the the probe, for sure.

If you can't go over 100mA since you fried/replaced R55 and you are sure about the probe accuracy, there's some other components bad, so I'm with astronomicum be careful before applying power to the amp, do you have a light bulb current limiter?

By placing a 1 ohm 1W 1% resistor from the power tube cathode to ground you can probe with your multimeter how much mV you see across that resistor, by ohm law it will be the same as the amount of mA.

[EddyMunster]

I've never heard of light bulb current limiter sadly. Thanks for the suggestion and for the explanation on the resistor. I checked the bias probe against a Marshall haze 15 amp. I was able to bias it to 25mv each side using the bias pins and ended up with 10 watts per 6v6 tube on the probe. I didn't write down the ma value but I do remember it was very close to the 25mv except in ma. I wanted 70% range for the tubes and that's where the reading put them. Looks to me the probe is fine. Does that make sense?

[DeepBias]

A light bulb current limiter is a safety tool that uses an incandescent light bulb to protect electronics from damage caused by short circuits or faults. You can buy one on ebay but it's simple to build yourself, it's just a little setup to fit a traditional incandescent light bulb in series with the "hot" from the wall 120VAC outlet so the bulb's resistance limits the flow of current. Check on google and you'll see a lot of diagrams, very cheap and useful to avoid damage.

Look inside you Haze and you'll see those 1 ohm resistors to probe bias.

[EddyMunster]

That sounds like an interesting gadget but I really don't know what to do at this point.

[glass54]

Eddy
See extract from Sluckeys Scrapbook Plus link to his site.
A lot of documented knowledge there 
https://sluckeyamps.com/misc/Amp_Scrapbook.pdf
https://sluckeyamps.com/index.htm
Regards
Mirek

[EddyMunster]

Thanks Mirek! Looks like a pretty cool light bulb gadget. I will put one together and see what happens.

[EddyMunster]

Okay, I put together a "light bulb current limiter" with a 200 watt bulb. I watched a couple of simple videos on youtube on how to use one with a tube amp. Thanks you guys for that! That could have saved me a lot of fuses over the years LOL!

I tested both amps with it and got the same results on both. see pics

The first pic is right after turning on the power switch. The second pic is after a couple of minutes. The third is after turning on the standby switch. Same thing with the other amp.

[astronomicum]

Looks normal. With the amps plugged in without the LBL, and biased to 100mA across the standby switch as Carvin specifies, what are the voltage readings on both sides of R57? Also report again what your bias probes are reading.

[BrianS]

What problem are you having that would make you think that the Carvin spec would be wrong? 

[EddyMunster]

astronomicum,

On the first amp with the fluke I got 102ma across the standby. At R57 I got 471 vdc on one side and 464 vdc on the other with the fluke. With the pro1 bias probe I got a reading on V5 tube: 465 plate, 3 watts, 8ma. I verfied previously that the tubes are a matched set so the others are similar.

On the second amp with the fluke I got 103ma across the standby. At R57 I got  472 and 465. Using the bias probe on V5: 467v, 4w, 8ma. Also verified as a matched set previously.

[BrianS]

So the amps have little or no output?  Severely distorted output?

[EddyMunster]

Hi Brain,

Long story.....but here it goes

On the first amp I was following a procedure from "Eurotubes" for 100 watt Carvin amps using their bias probe pro1 with a new set of EL34's. By doing so it ended up with a burnt R55 resistor.

The resistor was cooked because when I first set the bias I had to crank the bias adjustment to max and it was still a little low using only the bias probe (around 30ma per tube or something). It didn't seem right that the bias adjustment would be maxed so I got out the fluke to make a comparison across the bypass switch (it was well over 200ma). I turned if off right away and noticed R55 resister was cooked which I later replaced. That's what got me started down this rabbit hole!

I also found and looked over the attached bias procedure from Carvin (look about page 32 for some numbers). It seemed like a good a sanity check.

I called one of the guys at Eurotubes who walked me through a few troubleshooting steps and he said I should see only about 30ma on top of the total of the 4 power tubes across the standby. So at say 165ma at the standby the power tubes should be around 34ma each. At the end of the day he said I have some issue with extra current from somewhere.

I could not find a shop interested locally to tackle it so I picked up another L3 amp to do a comparison. I figure I can sell it if it's okay or fix it then sell it if not.

The second amp was set at 160ma when I got it. I noticed the same resistor is a little cooked but not as much as the first amp I cooked! The resistor checks out so it's still good. I set the bias down to 100ma. The second amp has similar readings just like the first amp. Both amps approx. 8ma per tube when set at 100ma across the standby switch!?!?

I thought to myself, "unlikely I have 2 faulty amps or is this just the design?" So I reached out to this forum for some answers.

They do sound okay to me and get plenty loud.

Bottom line is: I just want to know if I have 2 amps that really have an issue or are they working as designed?

[astronomicum]

astronomicum,

On the first amp with the fluke I got 102ma across the standby. At R57 I got 471 vdc on one side and 464 vdc on the other with the fluke. With the pro1 bias probe I got a reading on V5 tube: 465 plate, 3 watts, 8ma. I verfied previously that the tubes are a matched set so the others are similar.

On the second amp with the fluke I got 103ma across the standby. At R57 I got  472 and 465. Using the bias probe on V5: 467v, 4w, 8ma. Also verified as a matched set previously.

Based on the schematic, the HT current has two paths after the standby switch. It either goes to the OT center tap (then on to the power tubes), or through R57. We can calculate the current through R57 because we know the resistance and the voltage drop. For the first amp, that is (471V-464V / 360ohms = 0.019A or 19mA. There is 102mA through the standby so the current to the OT CT is 102mA - 19mA = 83mA (about 21mA per tube), not 32mA as your bias probes would have you believe, assuming of course that the 102mA measurement is accurate. I believe it is because the voltages at R57 are very close to those specified on the schematic. If the current to the power tubes was in fact 32mA, the voltages at R57 would most likely be higher than expected. The second amp yields similar results. IMO, the amps are operating with a cool bias but as they were designed. I can't say what is happening with the bias probe but sure does appear something is amiss.

[EddyMunster]

Sorry a little over my head but are you saying the bias probe may be faulty since it is only reading 8ma per tube? I did verify the probe on a Marshall amp with 6v6's. Could the EL34 readings be off yet the 6v6 readings be correct?

[BrianS]

I guess what I would do at this point would be to:
-Replace the burnt up resistor(s)
-Make sure the output tube type switch is set for whatever tubes I have installed (EL34 or 5881).
-Hook up an 8ohm resistive load.
-Adjust the bias pot per Carvin's instructions on the schematic:  100mA across the STDBY switch.
-Inject a 200mA, 1khz sinewave into the input jack
-Set amp to the "clean" channel
-Adjust Vol./Master controls to see if you can get a clean sinewave output at around 100 watts RMS.  If my math is right, that should be around 28V RMS across your 8ohm load.  I typically measure output power with an O-scope and do the math, but you should be able to measure with your Fluke 87 and get the RMS voltage.
Doing this exercise will help confirm your statement  "They do sound okay to me and get plenty loud."  Keep in mind though...30 watts of output power is LOUD, but still not right if the amp is supposed be 100 watts.  This is why measuring is more important that hearing.
If you get some numbers that give you significantly less than 100 watts of output power, double check your set up and procedures.  If low output persists, then you'd have to move on from there to find the problem.
I can't say for certain why your measurements using the bias probes is off.  I've never had one of these amps on my bench, and I don't use the same bias probes as you.  However, the specs and procedures outlined on the schematic should give you a good result.
Do you have a signal generator?  How about an oscilloscope?

[EddyMunster]

Unfortunately I don't have a signal generator or oscilloscope and it's outside of my skill level.

[EddyMunster]

astronomicum,

I ordered a basic socket bias probe on friday to use with my fluke meter that will give me a direct comparison with the Eurotubes Pro1 bias probe. The Eurotubes probe is the one that calculates the wattage in real time. It's a nice one. The basic socket probe will be here by saturday. That should give me more confidence in the Pro1 probe or prove that it's wrong.

[Willabe]

200w bulb is too big. Too hard to see the change in brightness.

Use a 100w bulb for a 100w amp. 100w to 60w bulb for 40w to 50w amps. 20w amps use the 60w bulb.

Lower wattage bulb gives better short protection.
 

[EddyMunster]

All,
   Got the other bias probe today and got no reading until it dawned on me my meter was set on AC current. Once set on DC current I got a reading that matches the other bias probe. I'm really happy the probes are working just fine but unhappy with the tech(me).  I feel pretty dumb and embarrassed about it. Also, it dawned on me that I've been measuring across the standby switch with my meter(s) set on AC current too! I didn't know there could be readings both AC and DC ma's. Sorry about the rabbit hole and I do appreciate everyone's time. I've learned a lot! 

Anyway, I now have 100ma "DC" across the standby switch and 21ma "DC" at each tube.

Since I did get some cooked resistors I am a little worried about raising the bias at this point. Any thoughts on that?

[DeepBias]

... until it dawned on me my meter was set on AC current ...

Since I did get some cooked resistors I am a little worried about raising the bias at this point. Any thoughts on that?

Glad to see you’ve tracked down the problem, it happens to the best of us 

Filter capacitors are rated at 450/500V. If you haven’t done it already you can do a visual inspection to see if you can find other suspect resistors.

The amps are sounding good ?

Seems you have a Carvin amp to sell...
 

[EddyMunster]

Good morning,
The only components that look a little cooked are R55 and R56. Everything else looks good. When I got the 2nd amp it was biased to 160ma but I was using the meter set on "AC" current which means it was really running at about 100ma "DC" current. Are the 2 resistors underrated at 1 watt?? I know from the 1st amp when I biased it up around 250ma AC (I'm guessing 190ma DC) and the tubes were about 30ma on the bias probe, that cooked R55 on that amp. If I was to change out the 2 resistors to 2 watt or a little more would that allow me to get the amp closer to 60% dissipation?

Right now I'm getting readings 102ma "DC" across the standby and using the bias probe, 462v, 9 watts, 20ma on each tube. 15 watts would be 60% but since my initial "AC current readings were 60ma higher than if I was reading "DC" current then those resistors would definitely cook.

What are your thoughts on that?

[DeepBias]

Hi Eddy,

Did you replace the cooked resistors ? Otherwise, it's better to do it before anything else.

100mA across stanby switch means each tube should be drawing 25mA, for a dissipation of 462V * 0,025A = 11,55W

If you bias each tubes at 15W dissipation they should be drawing ~32mA DC each (15W / 462V = 0,032A), for a total of 128mA DC across the standby switch.

You could try to increase the dissipation little by litlle and check if everything's correct, verify resistors temperature between each step.

Be sure to discharge filter capacitors before playing inside the amp ->

Carvin Legacy 3 bias ->

 

[EddyMunster]

Thanks for all the input. I did try going over 30ma before I posted the questions and ended up frying R55 resistor. R56 looks a little cooked as well. I had over 200ma across the standby at that setting and P11 was maxed out. That is why I picked up another identical amp in order to do a comparison. I have since replaced R55 with a 2watt 4.7k I found locally. I need to be sure it's either a problem that needs fixing or is it just how these amps run by design. I can't run the bias past the 100ma for now.

Thanks for the videos. As you can see above I did replace a resistor.

How do I go about checking the temperature of the resistors?

[EddyMunster]

I have a raytek. I will try using that and let you know

[EddyMunster]

I took a temp reading on R55 (new 2watt resistor) and R56, which are both close together on the board, with the Raytek and got 238°F when set at 102ma bias across standby. I waited about 10 minutes between readings. I slowly raised the bias to 167ma which gave me 445v, 15w, 34ma with the pro1 bias probe and 37ma with the 2nd bias probe and fluke meter. This is a different set of tubes so the bias P11 did not max out this time. I did not see much difference in temp on the resistors between the two bias settings. Is that too hot for these resistors?

I did scan around with the Raytek and saw some others, R103, R104, R108 at 164°F but most of rest of the board was around 100°F - 120°F.

What do you think? Should I change out the resistors to a higher watt rating?

[AlNewman]

I'm not sure the best way to measure temperature of a resistor, or what a healthy temperature would be.  But you can measure wattage or power across a resistor.  Not sure of your particular voltages, but if you look at your schematic:

R55, 4700 ohms= 415V on one side, 385V on the other.  That's a 30V drop.
30V/4700 ohms = 0.0064 A.
0.0064 A multiplied by 30V difference = 0.192 Watts.  A 1 watt resistor should be more than enough.

R56, 2200 ohms =  445 on one side, 415 on the other side.  Again, a 30V drop.
30V/2200 ohms = 0.0136 A.
0.0136 A multiplied by a 30V difference = 0.408 Watts.  Again, 1 watt resistor should be enough. 

Notice these resistors come after the power tubes, so there won't be much difference in current draw depending on how hard the amp is working.  Also, if you have much higher voltage differentials that what is noted on the schematic, for example enough to fry a 1W resistor, there is probably something else wrong with your amp.  Installing a higher rated resistor will likely only damage something further down the chain.

[EddyMunster]

Thanks for the formula......so I got 79 vdc drop across R55 and 44.5 vdc across R56. That's 1.32w for R55 and .99w for R56 right?  I let it idle for about 1/2 an hour at 15w at each el34 tube and got a high temp of 256°F at the resistors location.

I got approx. the same results on the other L3 amp. Definitely not the 30 vdc drop on the schematic. Both resistors on both amps are a little discolored but are measuring the resistance values shown on the schematic.

I don't see any other hot spots with the Raytek just the 2 resistors on both circuit boards.

What do you think? Why would there be a bigger voltage drop?

[AlNewman]

Did you measure the resistances of the resistors to see if they're in spec?

Those numbers point to something drawing higher current (an extra 10mA) from the D rail.

[AlNewman]

Have you replaced C81 and C82 yet?

[EddyMunster]

Yes, the resistors are in spec on both amps. I could read C81 on the board but I had to pull C82 to get a reading with my fluke 87v. They are both good on the one amp. I did not check the other amp.

I don't know what the D rail is. What is it?

[AlNewman]

Everything that's fed from this node.

[EddyMunster]

I'm not very savvy at reading the schematic. So the D node is connected physically on the board to any place on the schematic with the D symbol you circled?

[DeepBias]

I'm not very savvy at reading the schematic. So the D node is connected physically on the board to any place on the schematic with the D symbol you circled?

Yes

If you have turned the bias potentiometer all the way up, it is possible that the two filter capacitors C81 and C82 have been damaged, partially shorted filter capacitor(s) may draw extra current through R55/R56. Replacing them is a cheap insurance. After replacing check voltages, if abnormal, the issue is downstream.

You can remove preamp tubes one by one to see if heating disappears.

Do not just replace R55/R56 with higher wattage, it will mask the problem.
 

[EddyMunster]

Thanks for the D rail explanation. Does it matter the direction the D is pointed? I'm not at all familiar with that symbol.

As for the tubes, it was recommend I pull 3 of the four 12ax7's and to leave the inverter tube in place and check the voltage drop. I did that and got 63 vdc (79 vdc with all the tubes in). I also tried a completely different set of 12ax7 tubes. Got the same results on both amps, 63 and 79 at 100ma across the standby.

How would I check a cap, before swapping it out, besides using the capacitor function on the fluke 87v?

[shooter]

How would I check a cap, before swapping it out, besides using the capacitor function on the fluke 87v?

gatorclip a 100k across the cap, wait a couple minutes, use a 2nd gatorclip to  short across the cap, remove the 100k n clips, leave the shorting clip in place.  set your fluke to Resistance, + to cap +   - to cap -


remove the shorting gatorclip...LEAVE meter leads on cap
the meter should start running to infinity pretty quick


if it stops < 500k I typically replace

[AlNewman]

There's also 4 high voltage transistors that draw from that node, via the reverb and effects loop circuits.  So, if any of those were shorted that could also account for the extra current draw.  But the filter caps would be the first culprit to look for.

[DeepBias]

In addition to everything mentioned above, here are the common states in which a capacitor can be :

Good capacitor: The reading will start low and then steadily increase, moving toward infinity (OL) as the capacitor charges from the meter's internal battery.

Partially shorted/Leaky capacitor: The reading will rise but stop at a resistance value well below infinity. The lower the final resistance value, the leakier or more shorted the capacitor is.

For exemple ~1M (or higher) is ideal / excellent with very low leakage. ~550K is acceptable but aging.

Completely shorted capacitor: The reading will stay at or near 0 Ω (zero ohms).

Open capacitor: The reading will immediately show "OL" or infinity and not move.

Loss of Capacitance (Low Capacitance Value). Power supply ripple increases, unstable circuits, or weak signal coupling.

Please never forget to safely discharge the capacitors and mesure the voltage just to be sure, this is a must be done first and foremost.
 

[EddyMunster]

Thanks for all the help and suggestions! Greatly appreciated

I finally got some time to take an amp apart and removed C81, C82, C83 and C84 filter caps. I got well over 1megohm and was still slowly climbing on each cap using the method above.

Used a mesr-100 meter and they all check good.

Used the fluke 87v cap test and they measure very close to spec.

Used a VC60B+ insulation tester (first time) but the manual sucks and it took me forever to figure it out. I got 75+ megohms at 500vdc on C83-C84 and over 200 megohms at 250vdc on C81-C82. Stuart UKguitarampguy on youtube has a video using an old analog megger for that method of leakage test. The VC60B+ is digital and reacted slower but it seems to work.

Anybody use one for capacitors or anything else? With my limited knowledge in mind, they seem to be fine. What do you guys think?

**Link the megger video**   

&t=85s

[AlNewman]

They're probably ok, but if they're out it isn't a big deal or expense to change them.

With the amp off and discharged, I would check resistance across all pins of Q3, Q4, Q5, and Q6.  You should be able to do this in circuit.  If they are gone, they will measure as pretty much a dead short. 

Another way is to measure the voltage drop across R98, R108, R34, R104 and measure the current draw using the same formula as you did for the dropping resistors.  There shouldn't be much more than about 10V across any one of them.

[EddyMunster]

AINewman, thanks for the help!
   
I checked Q3-Q6 and none were shorted.

I am seeing a voltage drop across R98, R108, R34 and R104 of 148 vdc, .46 watts, 47k 1 watt resistors.

I also measured R13 and got 148 vdc drop, .09 watts  220k 1/2 watt resistor.

R10, 156vdc drop, .1 watt   
R3, 198vdc drop, .17 watt
R31, 194vdc drop, .17 watt   all 220k resistors 1/2 watt

[AlNewman]

The plate resistors on the tubes seem to be ok.

But the supply for the transistors are drawing twice as much current as the whole circuit should.  I'm kinda thinking all those transistors might be fried.  I don't see a single part that would cause all of the transistors to draw that much current, and I can't think of a reason why all of them would go at once.  Somebody else might have some ideas.