5E3 Build Smoke Comes out of Resistor

[apeontheweb]

I built a chinese Ebay 5E3. I plugged it in and it sounded great. I smelled smoke and shut it down. The smoke came from a 4.7K voltage dropping resistor in between the first and second filter capacitors. On some schematics, it's noted as a 5K resistor. To help orient you on the schematic, it is tied to a 22K resistor I was wondering if anyone could help me troubleshoot my problem. Edit: I realized when I the resistor started getting hot that I was playing with the amp in the orientation so that the open side was laying down. There was no ventilation. Could that maybe have caused the resistor to overheat?

[Banjan73]

My first thought is that the resistor has too low wattage. But since this is a kit... I don't know.
Have you measured current/voltage drop over the resistor (if it still measures 4.7k)?

[JPK]

This layout shows it as 2 watt. I checked a Weber and Robinette layout and they're all 2 watt. You can see them on my populated, un-soldered board on the bottom left. Did they supply a lower watt resistor? Got a BOM or pics of it?

[apeontheweb]

The little bag the resistor came in said "2 Watt 4.7K" and I measured it and it was 4.7K. I don't know maybe it wasn't a 2 Watt but how would I know? The resistor still reads 4.7K. What voltage drop might i expect to see over it? (Edit note: I originally wrote 47K here and it was a typo. I went back and corrected this to read 4.7K)

[SEL49]

The little bag the resistor came in said "2 Watt 47K" and I measured it and it was 47K. I don't know maybe it wasn't a 2 Watt but how would I know? The resistor still reads 47K. What voltage drop might i expect to see over it?

Should be 4.7K, not 47K.

[stratomaster]

Something at that node is drawing enough current to kill that resistor.  I'd be inclined to think it's the filter cap for that node, but it could be anything there.

[JPK]

Should say 4k7. Not 47k. Should read close to 4.7k. If they actually supplied a 47k that's not gonna work. The other one needs to be around 22k. Both 2 watt as my layout shows.

[apeontheweb]

Yes it is a 4.7K not a 47K. I also verified with my multi meter. I apologize for my lousy communication when I accidentally said a 47K resistor.

[JPK]

Was it 2 watt? It should look like it. Much bigger. If not look at the tube socket wiring. Make sure it's all correct.

[apeontheweb]

Yeah the resistor is a bit bulkier than the others.

[AlNewman]

Sounds like a good time to build a light bulb limiter.

[tubeswell]

Something at that node is drawing enough current to kill that resistor.  I'd be inclined to think it's the filter cap for that node, but it could be anything there.

^This^


The reason the resistor is smoking is its drawing too much current.

The two pathways that draw current from that resistor are the 6V6 screens and the preamp B+ supply.

Therefore, assuming the amp was built properly, there could be either a failure (short) across the screen node filter cap, or a short on one of the 6V6 screen grids that is dumping excess current to ground (causing all B+ voltage to be dropped across that resistor). Or it could be that you have a short to ground behind the board at that filter cap's power supply node.  If the screen filter cap voltage node measures '0VDC', then it's most likely one of this possibilities.

Please measure the PS rail voltages on all filter cap nodes and report back

[Banjan73]

As said by tubeswell: current is too high.
I've experienced this myself, and then there was a short to ground.
Another longshot thing is that resistors actually have voltage ratings. Since this is in the B+, they should have voltage ratings well above the actual voltage you have.
Since this is a chinese thing, I may question the quality of those resistors (rating wise). But hopefully, they have considered this.

[apeontheweb]

I'm pretty sure that 4.7K resistor is a 2 watt -- some folks were asking. I attached a photo. The red arrow on the photo points to the resistor that heated up. I am going to get a replacement resistor today and then begin to test the circuit. I'll report back. Thanks so much.

[tubeswell]

Further to my previous post - some possible fault pathways

[apeontheweb]

Here are some numbers I can share with you.
B+1 347 VDC
B+2 300 VDC
B+3 237 VDC
Mains amp draw: 0.55A
Voltage drop across 4.7K 2 Watt resistor with guitar plugged in and volume cranked 109VDC at its peak but hovered around 90VDC. (Edit: I calculated that at 90VDC that's right about 2 watts and 109VDC is 2.5 watts!)
I think these numbers are all normal?
I wonder if the resistor overheated because I was testing the amp with the open side faced down? No ventilation. I was playing with the amp cranked. Also I was using a 16 ohm speaker because Chatgpt said that wasn't a problem. But I've since read from Weber speakers that you must use an 8ohm or 4ohm load for an amp that wants to see 8 ohms. But that is an aside. What do you think is there any evidence pointing to why my resistor overheated? Could be an intermittent problem as well.

[apeontheweb]

Hey Tubeswell, Thanks for the help. I'm sorry I didn't get all the values you were asking me to get but here's a few. I had a variac and current limiter in the circuit when I did these tests. They were done at 110VAC.
5Y3 tube socket:
P2 : 275VDC
P4: 262VAC
P6 : 262VAC
P8 : 267VDC
Please also see my above comment where I listed some other specs. Any thoughts? I'm feeling like everything looks good? Maybe an intermittent problem? I dunno.

[tubeswell]

The 5Y3 voltages are a start (and you should measure all pairs of pins in VAC (i.e. Pins 2 and 8, and Pins 4 and 6) but knowing the VDC of Pin 8 is handy because its gives us the B+ voltage.


But you need to get the DC voltages at each filter cap node in the power supply, i.e.,


Reservoir cap =
Screen supply node cap =
Preamp supply node cap =


The reservoir cap voltage should be the same as the VDC on Pin 8 of the 5Y3.


The screen node should be about 60V lower than the reservoir cap because we would expect the 4k7 resistor to be 'seeing' about 14mA at idle (if the amp is operating normally). 60V x 0.014A = about 1W (so a 2W resistor should be fine).


The preamp node should be lower still.

[SEL49]

Voltage drop across 4.7K 2 Watt resistor with guitar plugged in and volume cranked 109VDC at its peak but hovered around 90VDC. (Edit: I calculated that at 90VDC that's right about 2 watts and 109VDC is 2.5 watts!)
I think these numbers are all normal?

These wattage numbers explain why that 4.7K is discolored. The general rule of thumb is to double the actual wattage dissipation of the resistor then choose the next higher common wattage rated resistor. Based on these numbers, that 4.7K should be rated for 5 watts.

However, your B+ numbers indicate there is only 47V dropped across that 4.7K at idle. This is a more typical voltage drop. It means that the calculated power dissipated by the 4.7K (at idle) is only .47W. A 1 watt resistor should be sufficient.

The problem... Your 6V6s are drawing excessive current when the amp is cranked. Fender did not use any screen resistors and I suspect your Chinese clone didn't use them either. The solution... If you are gonna crank this amp you should really put some 470Ω/1W in line with each screen. 1K/3W would be even better.

EDIT... Look at this layout to see how to easily add screen resistors to your amp.

https://sluckeyamps.com/5e3/5e3.pdf

[apeontheweb]

Tubeswell- Here are some more numbers. These were all taken at 123VAC from the wall and no current limiter in the circuit. Thank you for any insights.
B+1 347 VDC
B+2 300 VDC
B+3 237 VDC
Mains amp draw: 0.55A
Voltage drop across 4.7K 2 Watt resistor with guitar plugged in and volume cranked 109VDC at its peak but hovered around 90VDC. (Edit: I calculated that at 90VDC that's right about 2 watts and 109VDC is 2.5 watts!)
I think these numbers are all normal?

[apeontheweb]

Hey Sel49 - screen resistors sounds like a good idea. well, I'm not really one to judge a good idea from a bad idea as this is my first build. But not melting another 4.7K resistor sounds like a good idea at least. But let me ask you, how do the screen resistors change the sound of the amp? Maybe they make it a little quieter? Is there a tonal change? Thank you.

[acheld]

You can trust what Tubeswell and SEL49 are saying.  These guys have been around a long time and consistently provide great advice.

As to your earlier question about lack of ventilation causing the magic smoke to come out -- no it won't. (It can cause problems in the longer term, but not this.) 

And no worries, this has happened to all of us here.  The trick is having the patience to work it through.   And no, ChatGPT will not help you this year.

[Williamblake]

It has been said several times already: If a resistor gets hot: Measure the voltage across that resistor and calculate the current going through that resistor.

[apeontheweb]

Here are all the voltage readings I've done so far. All readings at 123VAC wall. No current limiter on the amp.

6V6 :
P2: 19.46DCV
P8: 19.46DCV
P4: 318 DCV
P6: 318 DCV
P2-4: 3.05ACV
P4-6: meter didn't settle on a value

6V6 Plate Current: 34.6 mA
Plate dissipation: 11.0 W per tube

B+ DCV
B+1 : 347 DCV
B+2 : 300 DCV
B+3 : 237 DCV

Mains amps: 0.55 amps

4.7K Dropping resistor when amp at full blast :
109 VDC peaks (2.5 watts)
~90 VDC average ~(2.0 watts)

Anything interesting here?

[SEL49]

Anything interesting here?

Hello. Are you not listening?

[apeontheweb]

Yes Sel49 I'm listening. If you scroll back you'll see I already replied to you. I've been gathering numbers because another user has asked me to.

[SEL49]

My apologies for sounding a bit blunt. I did see your reply, but it seemed to me that you were circling around. Just trying to keep you focused.

[tubeswell]

6V6 :
P2: 19.46DCV
P8: 19.46DCV
P4: 318 DCV
P6: 318 DCV
P2-4: 3.05ACV
P4-6: meter didn't settle on a value

6V6 Plate Current: 34.6 mA
Plate dissipation: 11.0 W per tube

B+ DCV
B+1 : 347 DCV
B+2 : 300 DCV
B+3 : 237 DCV

Mains amps: 0.55 amps

4.7K Dropping resistor when amp at full blast :
109 VDC peaks (2.5 watts)
~90 VDC average ~(2.0 watts)

Anything interesting here?

Thanks for these. These idle voltages are normal


The problem (as you have surmised) appears to be excessive output tube screen current when cranked. As has already been noted, screen grid resistors will help limit peak screen current. Try 470R in series between the B+2 node and each 6V6 screen grid socket pin. This should help decrease peak current through the 4k7. (But you should also replaced the cooked 4k7)

[tubeswell]

As to the higher mismatched load, yes that will cause excessive g2 current (especially if your tubes are biased near or at Pmax). plugging a 16R speaker into an 8R OT tap will mean your 6V6s are seeing a 16k plate-to-plate load, which is 1/2 the gradient of an 8k loadline. This means that the loadline will be well under the knee of the Vg0 grid curve, and so on the part of the signal cycle when the plate voltage drops, the screen current will rise more rapidly than normal. Running a proper load will reduce peak screen current.

[apeontheweb]

thanks for your help on this SEL49 and Tubeswell and all. I appreciate your time!

[apeontheweb]

I installed a 470 ohm 2 watt resistor on each 6V6's screen grid. I measured the voltage drop across the 4.7K resistor that had heated up. I did a test right before adding the screen grid resistors and I did a test after adding them. It's a little tough to tell exactly the voltage drop difference -- but I believe it's between 5 and 10 less DC voltage drop after the two screen grid resistors. I might have to bump these resistors up. Can anyone surmise what the change in the amp sound might be by adding the resistors? Quieter? "Sweeter" overdrive?

[SEL49]

It will sound more reliable. You're the only one that can hear the amp. Do your ears think the sound has changed?

[Rontone]

how do the screen resistors change the sound of the amp? Maybe they make it a little quieter? Is there a tonal change?

I have tried 470R, 1K and 2K for screen stoppers on a 5F1 champ with no real change in tone or "breakup" levels, you will be fine with those 470R in there

Mine also had the speaker attached on the wrong output transformer tap when I bought it, I wondered why it was running so hot, I checked everything and didn't consider that the speaker was on the wrong tap....

[apeontheweb]

Thanks. I could not hear any change with the 470 ohm resistors on the screen grid. But I was asking how the mod effected the amp sound because I want to use bigger resistors and wanted to know how that might change the sound. There's a different mod on the rob rob site called a phase inverter grid stopper resistor mod. When I was asking if the screen grid resistors would sweeten the overdrive tone, I think I was mixing up that phase inverter mod with the screen grid mod. Anyways, I think I'll try bigger resistors to see if I can lower the voltage drop across the 4.7K resistor some more. Thanks again, everyone.

[SEL49]

Many people will use 1K screen resistors.

[tubeswell]

I installed a 470 ohm 2 watt resistor on each 6V6's screen grid. I measured the voltage drop across the 4.7K resistor that had heated up. I did a test right before adding the screen grid resistors and I did a test after adding them. It's a little tough to tell exactly the voltage drop difference -- but I believe it's between 5 and 10 less DC voltage drop after the two screen grid resistors. I might have to bump these resistors up. Can anyone surmise what the change in the amp sound might be by adding the resistors? Quieter? "Sweeter" overdrive?

470R screen grid resistors won’t change the way the amp sounds, but they will help eat up some of the peak screen current when the amp is driven hard.

[Meteorman]

It will sound more reliable.

 
Beautiful!... that made me laugh out loud