Noise and Excessive Gain, Hoffman Pro Jr Board in a Carvin Vintage 16

[CHEZ]

I bought my Carvin Vintage 16 amp 8 years ago and after much enjoyable use the PC board failed. I decided to do a Hoffman Pro Jr conversion. This is my 4th amp build. Both amps are SS rectified, have two 12ax7s and 2 EL84s, and an 8 ohm 12" speaker. The PT looked close. I have completed the build, double checked the connections and the amp works. Problem is excessive gain and noise (hiss and hum) even when nothing is plugged in and the volume down. The noise changes some as I increase volume and tone but does not get louder. I have been playing electric for 50+ years and this amp does not sound right. I have to dial the volume back on my guitar to a 2 to have something like a normal level at the amp's lowest possible setting, then there is the noise. It is not HORRIBLE but annoying. I swapped out tubes, identified a microphonic V1, with only marginal improvement. Voltages are on the high side, Pro Jr wants a B+ of 325vdc, I am reading 370. I have tried larger resistor values on R31, up to 2Mohms, bringing the B+ down 352 or so with no improvement in sound. I am able to dial the bias to the required -10.4vdc. I have studied the info on Hoffmanamps.com and related topics on this forum without finding solutions. I will put up pictures. I grounded the input jack directly to the chassis since these pictures were taken. I would greatly appreciate suggestions.

[Latole]

I don't like those very long red wires and their routing. They may act like antenna for hum.
IMO the should run close to the chassis.
The wires should not be parallel to each other and they should cross each other at 90 degrees.

The V1 can give the impression of being microphonic simply because it is located at the beginning of the amplification chain. This is quite normal.
There is still a level of microphonic that may not be normal too

[CHEZ]

THANKS! I will address my leads to make sure they are short and not parallel. The 12AX7 I used to replace V1 was much less microphonic, leading to my assumption of a bad tube. I am putting up some more photos for you to see. Again, thanks!

[Latole]

Vintage Fender amp show nice wiring

[CHEZ]

That gives me a standard to emulate. BTW the only modification I have done to this amp is the variable negative feedback. Do you think my high B+ is a problem? The schematic shows 325vdc and with the stock components I have 370vdc. I have attached one more foto from the original batch.

[Latole]

It is difficult for me to say what is causing your noise.
There are a lot of details to check and without having the amp in front of you and using an oscilloscope.

The B+, this slight difference is not a problem.

[CHEZ]

That is a relief about the B+. Unfortunately I do not have an oscilloscope. I will get busy applying your suggestions. I also now have some different tubes to try, including 7025 which are billed as low noise. The outrageous gain bothers me more than the noise. It sounds loud, raw, and on the verge of breaking up even at the lowest volume settings, making the amp unusable for me. I played this same amp for years with a different circuit, but the same transformers and same tube combination. I think I must have a wiring error. I will run each connection and component down, again.

[EL34]

That gives me a standard to emulate.

Better yet, look at my original install pictures
https://el34world.com/Hoffman/Pro_Junior_Build.htm

[ac427v]

I would first review the grounding. I see a Green with yellow tracer wire (heater ground?) screwed to a chassis hole cover. That may not be a secure electrical connection. Did you also use 47 ohm heater resistors to ground? That would cause problems if duplicated by the transformer heater ground wire. I advocate choosing a known good grounding strategy and following it precisely. Changing R31 to 2M ohms will cause the B+ to discharge very slowly at shut down. Your power supply caps may still be partially charged when you open up the chassis. Be careful in there!

[uki]

I don't like those very long red wires and their routing. They may act like antenna for hum.
IMO the should run close to the chassis.
The wires should not be parallel to each other and they should cross each other at 90 degrees.

Make sure those red wires are no longer than it really need to be, keep then away from from the heater wires and touching the chassis. Also the white wires from the pots, it is a good idea to use a shielded cable.

A picture of the whole chassis to see the entire circuit would be a great help!

The outrageous gain bothers me more than the noise. It sounds loud, raw,

What about the bias, did you check it? Are the PA tubes red plating?

[sluckey]

The outrageous gain bothers me more than the noise. It sounds loud, raw, and on the verge of breaking up even at the lowest volume settings, making the amp unusable for me.

Disconnect the negative feedback (NFB) wire from the board. Any better? If so, swap the OT primary plate leads and reconnect the NFB wire. BTW, placing that NFB pot right next to the input jack is just begging for problems.

[CHEZ]

Thank you all for your time and kind help.
1. grounding; that ground you see as a chassis hole cover is the original ground post in the Carvin amp, it is integrated into the chassis. I have retained the Carvin primary through secondary pc circuit as it is switchable 110/220 fused, and switched from the factory and gave no problems. I have also retained the OT mini pc board that is switchable 4ohm, 8ohm, and built-in dummy load. Both of these circuits ground to that post. I have to assume it is a very good ground. I did install the two 47kohm resistors on the Hoffman board next to the rectifier circuit as the schematic and layout show, and that is currently grounded to that same ground post. Please let me know how to improve that.

2. The bias is dialed in at -10.4vdc, the tubes are not red-plating.

3. The white wires you see going to a pot were the variable negative feedback mod I installed. I have now disconnected it and returned every component and connection to Hoffman and schematic and layout specs including a 3 watt 470k resistor at R31.

I checked voltages again. B+=352vdc, Y=322vdc, Z=259vdc, C-=-10.4vdc. I have shortened the leads where possible. The problem remains unchanged. I can best describe it as sounding kind of like a window air conditioner running... a blend of hums and hisses. With the volume turned all the way down the hum (not low like 60hz, but higher) is a bit louder, and the other noise is almost gone. At a certain place on the volume control path the hum is gone but the other noise is louder. I am attaching 4 current pictures.

Also please note that when I first powered up this amp using my current limiter (which I discovered I had wired the bulb in series with the neutral, not the hot ) a curl of smoke rose from R32. I instantly shut the amp down and found I had allowed R24's lead to touch the common lead underneath. I now have a piece of gorilla tape to separate them. I tested components, replaced C12 and D2, rewired my bulb limiter. And everything works as is does now.

Thanks again.

[sluckey]

OK, so the NFB pot is disconnected and you reconnected the feedback wire as per Hoffman's layout. But did you do this...

Disconnect the negative feedback (NFB) wire from the board. Any better? If so, swap the OT primary plate leads and reconnect the NFB wire.

[Latole]

" 2. The bias is dialed in at -10.4vdc, the tubes are not red-plating."

Where do you reading that ?
 
To me, the useful bias reading is the tubes cathod milliamp
Milliamps X plate voltage = watts
Bias = 60 to 70% of tube power

[CHEZ]

I did find that the hiss was reduced by disconnecting the NFb. I switched the OT connections and there was some improvement. Now disconnecting the NFb adds hiss.
I am reading the C- where it is shown on the wiper of R29.

[Latole]

I am reading the C- where it is shown on the wiper of R29.

Where is R29 !
I did not see a link to a schematic or I'm blind ?

[CHEZ]

My bad, find it attached and please do help me get a true reading.
After changing the OT to power tubes wires around I took some time and played my electric guitar through the amp for about 20 minutes. I would say the primary issue at this point is the hum. The quality of the guitar sound is very usable, but it sure was a relief to finally turn the thing off!
Also I have heard some builders keep track of the polarity of the twisted pair of heater wires and make connections consistent with polarity. I didn't do that. Please comment

[EL34]

It sounds to me like you did not follow Hoffman's law.


If you wired the amp correctly, it would be working.

Did you read all of the assembly documents on the library page here?
Wiring, biasing and all that is covered in those documents.
https://el34world.com/schematics.htm#Hoffman_Pro_Junior_Conversion

[Latole]

I would never rely on the voltage of C- ( -10VDC)
C- is a simple command to the tube and not a result.

Depending on the tube and the plate voltage, the same C- will make it react differently. The tube will be biased differently
You have to rely (as shown in the schematic) on the 33 mv with a plate voltage of 320 volts

If your plate voltage is different, the -10 vdc will not be the right voltage.

Use my formula written previously
For example 33 mv X 320 V = 10,56 watts
EL84 max power; 12 watts
Best bias are 60% ( my choice ) to 70 % of this 12 watts

This amp is biased very very hot at 88 % !!!

 

[tubeswell]

1. grounding; that ground you see as a chassis hole cover is the original ground post in the Carvin amp, it is integrated into the chassis. I have retained the Carvin primary through secondary pc circuit as it is switchable 110/220 fused, and switched from the factory and gave no problems. I have also retained the OT mini pc board that is switchable 4ohm, 8ohm, and built-in dummy load. Both of these circuits ground to that post. I have to assume it is a very good ground. I did install the two 47kohm resistors on the Hoffman board next to the rectifier circuit as the schematic and layout show, and that is currently grounded to that same ground post. Please let me know how to improve that.

The way that many commercially-made amps achieve their grounding is more complex that just attaching everything to one grounding post. There may be isolation resistors and other signal ground separation mechanisms within the ground return path(s) on the Carvin PCB that you have not considered. These are not necessarily logically drawn on commercial schematics. (Peavey schematics are notable for this type of omission. Carvin schematics may be the same - I haven't looked)


Therefore, I would ditch your interpretation of the Carvin ground scheme and start from scratch (if it were my amp). YMMV
Merlin Blencowe's chapter on grounding is a good place to start in terms of setting out a proven signal ground layout. http://www.valvewizard.co.uk/Grounding.html

[CHEZ]

I will crunch the true bias numbers shortly. Tubeswell, you have correctly put your finger on my need for a "well grounded" education. Challenge accepted. I will study the grounding article, reread the Hoffman build documents and rethink my random grounding scheme. All my future builds will be the better for it. Thanks to all. I will report back with any progress, problems, revelations and supplications.

[tubenit]

Hoffman's grounding scheme has proven very helpful to me for having quiet amps at idle. I have tried other grounding schemes instead of Hoffman's with less success.


Grounds (el34world.com)



With respect, Tubenit

[EL34]

Hoffman's grounding scheme has proven very helpful to me for having quiet amps at idle. I have tried other grounding schemes instead of Hoffman's with less success.

Wow, I must have drawn that diagram way before getting a more modern image editing program
It works though

[thetragichero]

reminds me of macpaint like we had on ancient macs in graphic arts class

[EL34]

reminds me of macpaint like we had on ancient macs in graphic arts class

It was probably done on a windows 95 computer and MS paint?

[CHEZ]

Since the Carvin did not come with a fender style lamp holder, just an LED sticking up off the PC board, I had just soldered the 6.3v pair directly to an led lamp and taped it in place while waiting for the lamp holder to come in the mail. Today I installed it and found a bad splice on the twisted 6.3v pair. I got rid of the splice, shortened the connection and cleaned everything up. On powering the amp up I found the hum reduced by about 25%. The gain is now normal and the hum has gone from bad to not too bad. Still there with the volume down. I pulled V1 and the hum went down a little more. I pulled V2 and the amp became dead quiet.

[CHEZ]

I find that V4 plate current is 61 mA with 351 plate voltage giving plate dissipation of 21.37 watts.
V3 plate current is 42 mA with 349 v plate voltage giving plate dissipation of 14.74 watts.
If the EL84 max plate dissipation is 12 watts and our goal is 60% of that, 7.2 watts, then we need to get these numbers down.
Please advise.

[Latole]

Don't put power tubes in the amp for now .

[sluckey]

I find that V4 plate current is 61 mA with 351 plate voltage giving plate dissipation of 21.37 watts.
V3 plate current is 42 mA with 349 v plate voltage giving plate dissipation of 14.74 watts.

What is the voltage on pin 2 of each EL84?

I had just soldered the 6.3v pair directly to an led lamp and taped it in place

That will insure the LED dies as soon as you turn on the power.

[CHEZ]

Voltage on V3 pin 2 is -7.8vdc
Voltage on V4 pin 2 is -7.7vdc

The plate dissipation readings in the previous post and these voltages were taken with the bias trim-pot increased to the highest resistance. The lamp I connected is a 6 volt LED fender style amp bulb replacement. It has worked this whole time with no problem.

Thanks for your input and guidance! Where do we go from here?   

[Latole]

Voltage on V3 pin 2 is -7.8vdc
Voltage on V4 pin 2 is -7.7vdc

 

You need more bias voltage like -30 dc

I never saw a so low bias voltage, I don't understand this amp desing.
Do you use brand new tubes  ?

[sluckey]

Yes, EL84s need more negative bias voltage on pin 2, but not -30V! You need something between -10V and -15V. And you possibly need a fresh pair of EL84s.

That LED bulb has a built in current limiting resistor, so that's fine. I thought you just had a naked LED. Details are helpful.

[Latole]

R30 and / or R28 ; would need a higher value

[CHEZ]

By turning the bias trim-pot fully counter clockwise EL84 pin #2 now has -17vdc. The hum is almost gone and the amp plays respectably. Plate voltage, however is 375vdc.
Suggested values for those resistors?

[Latole]

By turning the bias trim-pot fully counter clockwise EL84 pin #2 now has -17vdc. The hum is almost gone and the amp plays respectably. Plate voltage, however is 375vdc.
Suggested values for those resistors?

Now check bias milliamp at EL84 pins 3 if close to 60-70% power
Before you have -8 volts bias, Now with -17 you should be ok

[Latole]

"...were taken with the bias trim-pot increased to the highest resistance."
-CHEZ


That is was not thru, it was with lower resistance

Bias trim pot with lower resistance = lower bias voltage = -8 volts
    "         "          highest   "           = highest voltage  = -17

[sluckey]

Suggested values for those resistors?

You can vary your bias voltage from -8 to -17. That's about perfect. Leave the bias circuit alone.

[CHEZ]

I have tried several combinations on those resistors, but the amp performs best with the original values. Hum and hiss are very low, amp plays and sounds wonderful...But this high dissipation is going to be rough on my  El84s. I have two pairs to choose from, both NOS vintage. I will order more to try the amp with a fresh matched pair.

[sluckey]

How are you measuring plate current?

[CHEZ]

With the amp off and the meter set on ohms I measure the resistance of 1/2 of the Output Transformer between pins 3 and 7 on each EL84.
I then turn the amp on and set the meter to dc volts. I measure each EL84 between pins 3 and 7 to get the voltage drop. I get plate current by that voltage by the resistance, PC=V/R. So my last reading on V4 was 374/.537M ohms or 5370 giving .069vdc plate current.

I then take the plate voltage between ground and pin 7 with the amp on, so 373vdc.

I get Plate dissipation using the formula (PV)(PC)=PD, so, 373x.069=25.7

[sluckey]

I think your method is flawed. Try this simple, easy method...

Measure the millivolts at pin 3. Since this is the voltage drop across the 1Ω cathode resistor, the current will be the same as the voltage. On Hoffman's schematic he shows 33mV on pin 3. That means there is 33mA flowing into the cathode. Now measure the voltage on pin 7. This is the plate voltage. Hoffman shows 320V on pin 7. Multiply the plate voltage times the cathode cathode current, ie 320 x .033. This is the plate dissipation. Hoffman's numbers give 10.56W.

This method ignores the screen current. This means that the calculated plate dissipation is actually a little bit more than the actual plate current. So the tube is actually running a little bit cooler than this calculation shows.

Hoffman included two 1Ω resistors (R98 and R99) just to make dissipation calculations quick, easy, and safe. Give it a try. Only takes a minute. Post your readings and calculations.

[CHEZ]

MV on pin 3 is .9 and 1.0 respectively.
Vdc at pin 7 is 374Vdc
formula yields 3.36 and 3.74 with an average of 3.5

So, I am running at 3.5 watts and everything is good, right?

[mresistor]

No  but Sluckey will be here shortly to tell you why.  How you are expressing Millivolts is wrong - 9 millivolts is  printed as .009 and we just say 9 mv.  If it was .9mv the tube would be almost shut off.
If you are running the tubes at 3 or so watts that is cold as ice . 70% of max plate dissipation (12W) for an EL84 would be 8.4 watts at idle. Also you take the readings with the volume pot at 0%.
You need to start turning the bias adjust pot towards less negative voltage,  It is helpful to use two meters .. one to monitor the cathode milliamps and the other to monitor plate voltage.. why.. because as you increase the current through the power tubes the plate voltage will start to decrease in relation to that current.  Then you have to recalculate the plate dissipation.  Do this in little steps while constantly checking the results and calculating the dissipation.

[sluckey]

I suspect operator error on the meter. Which meter? Show me a pic that clearly shows the whole meter, didplay dial setting, and probe connections.

[CHEZ]

The meter is a fluke 107

[sluckey]

You have AC millivolts selected!!! Put the meter on DC volts. That would be one click counterclockwise (left). Try again.

[CHEZ]

My bad

[mresistor]

so if your plate V is still 374 and you have .016 cathode current  you're running around 6W per tube which is around 50% of max plate dissipation disregarding the screen.

[PRR]

My bad

That always gets me too. Using one switch point for -either- AC or DC. I always get it wrong.

I used to have a meter which defaulted to DC, and I had to press the other button for AC. This made real trouble when colleagues borrowed the meter for AC power checks.

Now I have one which defaults to AC and I have to press a button for DC. Makes battery checks awkward.

IAC, it teaches to check ALL the settings along with the mystery-marks on the display.

[CHEZ]

Thanks everyone for your highly instructive input. You guys are awesome!