Pro Junior noise battles

[Murrayatuptown]

I'm working on a Pro Junior for a friend. Complaints are inferno-level EL84 bias, hiss and hum, and TBD wording about headroom and limited range of volume control before distortion.

I would like to ask some questions about trying to address ground loops but it's common to ask posters for schematics, gut shots, etc. The hiss and hum are not due to the mods performed, but they are in the photos...so I can't act like they aren't in my photos. My focus is on figuring out the grounding. The mods include things that aren't commonly done so I don't want to defend them. Happy to discuss them, but I'm not asking questions about them.

What vintage amp is it?
He thought it was a Series III. Main PCB is marked (c) 2002, which is pictured in a Fender layout drawing marked Rev. E. The schematic looks like Rev. E also. I have not been able to relate documentation revs to Series #, but that's not important to me.
The main PCB has a star ground system that relies on a steel screw and standoff connected to the chassis. I think this is consistent with later docs (probably Rev. E) as an earlier one showed a 'flying lead' ground with internal-tooth solder lug. The first 12AX7 cathode has a 47 uF cathode bypass capacitor and R6 was 22k, not 56k.

What's been done so far?

1) EL84 cathode current sampling resistors (1 ohm wirewound) were added on the tube socket board in series with the pin 3 cathode traces (cut) that go thru the ribbon cables to W1 to a local ground/return on the main PCB. Shielded twisted triplet wiring connects the cathode resistors and their common to isolated chassis-mounted tip jacks for DMM probes. A minor short ground loop was committed by connecting the low side of the 1 ohm cathode resistors together rather than running the ground all the way to the W1 local ground on the main pcb. That seemed like a worse thing to do. The shield is currently not connected at either end...

2) Adjustable EL84 bias via adding a chassis-mounted 25k pot (2-wire rheostat-wired) in series with R29 (15k for USA 60 Hz 120VAC). The 25k pot also has a 10k resistor in parallel for two reasons, limiting the range of bias voltage and open-wiper failure of a pot needs a different solution because the Fender bias circuit is not a straightforward series divider. Long story. It works, on the basis of grid voltage range (ideally I wanted it adjustable from the current voltage (-10.3 VDC or so) to something more negative.

3) Suggestions from a book author I'll leave unnamed were to try something very different than the method of handling the first 12AX7 gain and attenuation attenuation at the volume control and the tone stack. It was basically transformed into a tilt circuit with the signal level brought back to a higher level for s/n reasons and the 470 k resistor scaled down for noise reasons. To address the higher signal level there, an attenuator resistor was added across the tone stack to 'normalize' (back to where it was originally) the signal going to the LTP PI. The hope was that hiss and s/n would be improved in the first two 12AX7 stages (there is actually a 7025 there), and the tilt control offered more flexibility than the behavior of the original (in simulation). The gain and attenuation options in such a simple amp are limited, but what was done caused some concern (personal preference) in simulation. In practice, many people have observed the PI has a lot of gain, and people with noisy PJ's observe the noise is pretty obvious with the volume control at 0.

4) There is an added 'normalizing' resistor shown on a schematic as 6.2k is. The simulation files originally were left at 5.1k. So I started temporarily with a 3.3k resistor in series with a 5k trimpot, on a piece of PCB and twisted pair leads (still in lab rat mode). Hasn't been tested with an instrument yet. Trim pot turned down it's quieter, turned up it's noisier. That will get figured out, possibly as part of setting the volume control range.

5) Reduced Fender's R1 from 10k to 1k (to reduce Johnson noise).

6) Did several housekeeping things & experiments to attempt improving hum and noise.
Relocated the filament wiring by cutting the PT green leads shorter and attaching them to a terminal strip mounted where the AC cord safety ground is connected. Shielded twisted pair cable runs from the filament terminal strip to the socket PCB. The shield is grounded only at the end where the terminal strip is mounted at the AC cord safety ground. Removed the fixed humdinger resistors from the tube socket PCB and temporarily wired a 50 ohm wirewound hum balance pot off-board instead. Tried the AC safety ground point first - zero effect on him. Tried a filtered filament elevation (+24 V) at the hum balance pot CT. The hum was completely unaffected by the balance pot, with the CT either grounded or with DC elevation instead of a ground. Some experiments with where to ground the CT were tried. I concluded I wasn't (yet) dealing with a filament hum problem. I put the 47 ohm filament balance resistors in series with the ends of the 50 ohm pot. Since there is no improvement yet I don't need to waste so much current thru a 50 ohm pot.

Insulated the output jack which was formerly chassis-mounted. Removed the steel screw that grounds the main PCB to a steel threaded standoff. Insulated the PCB from the standoff with a 0.010" thick strip of Teflon and a nylon machine screw. Now there are no chassis grounds connections yet other than the AC ground which is also the shield/drain tie point for the filament wiring shield. I have verified a couple places in the amplifier do have local ground continuity, temporarily currently isolated from the chassis. I see Fender clearly has an intentional chassis ground on the schematic at the Input jack and the output jack, but they are very far apart via two separate connections. Good enough for Fender? Hmmm...I dunno. The plastic input jack chassis ground was obtained via the local ground network on the main PCB and the steel screw thru the PCB into the PCB standoff. The output jack was obviously chassis grounded. Before isolating those two grounds, I tried grounding some tube grids to the low side of R2 on the main board. Neither V1A nor V1B grid made any difference. Grounding either pin 2 grid at V3 or V4 produced near silence. I haven't tried the V2 PI yet...there are voltages I didn't want to short to ground. A capacitor was suggested. I'm reading Blencowe Designing Preamps for G&B Grounding chapter again.

Before I do that, I isolated the chassis grounds and started connecting chassis and PCB grounds together with clip leads to see if there were better or worse combinations. There were. I found one combination that was nearly silent but for an edgy whine.

All of this is with the amp chassis on a bench and a separate outboard speaker (old movie projector cab), so the rear panel with the foil shield is removed. I can see why that is essential with a chassis ground loop caused by two separate chassis connections. So I am not yet ready to go back to the two chassis ground points on opposite sides of the chassis.

I don't like the chassis ground 'diversity' via multiple chassis points, but I also don't like the idea of connecting the PCB input ground input all the way to the safety chassis ground with a long wire.

There are no chassis ground points using PT HV CT (there is none), B+ grown capacitor negative is drawn as a (power) ground, not a chassis ground. I tried initially to mount the filament wiring terminal strip at a power transformer bolt but there was too much varnish on the bolts/nuts...that's why the AC cord safety ground looked appealing (and there would only be ONE chassis ground point).

I am going to repeat some of this Saturday. The length and routing of the output transformer wiring and feedback wire make me wonder how hard it was for Fender to find that 'optimal' decision. Just a gut feeling.

[Murrayatuptown]

I am never going to get the attachments in the order I want for a single upload. The titles are mostly for my tracking, but I tried to number my actions and the photos similarly. Actions numbered 6.1, 6.2, etc. were after numbering MODS 1-5, and they are separate from and AFTER the mods were done. I initially therefore numbered them 0.1, 0.2, etc. but went with 6 to show the sequence.

I may try a different 12AX7 in the PI. It happens to be one of those Tungsol Russian reissues and they apparently generate a lot of opinions about hum. The 7025 is from The Tube Store. Since it's a modern source, I assume those are reissues/new construction. Hopefully they are made the same as the old 7025.

A few resistors were changed to RN55 type metal film as part of the evolution of this project. I don't think changing all of them is a meaningful thing...change the large values, as Johnson noise is proportional to (R)^0.5.

I see I ended up changing R28 from 56k CF to RN55 57.2k because I had that and not 56.1k. The simple change to R29 (add a pot) seems obvious now but I tried a couple of other very different approaches because I don't really like the original design is. I had a fully series divider at one point because I was going to try separate EL 84 bias pots. I decided that was not a good idea because it calls attention to so-called tube matching and it was better to not look at how closely tubes are matched on a fall Tuesday when there's a full moon.

The bias can be adjusted from about 65% of EL84 PD rating to worse than where it started. At 65% I can see little concern for matching. It's so much cooler now.

[Murrayatuptown]

I have it as quiet as I have ever heard it, but I have the 'audio returns' grounds connected by clip leads to the black wire of the output transformer secondary, at the now-insulated output jack. That is now my 'single point local audio ground'. The low side of R1 is connected to that with a clip lead because I isolated the (Rev. E) PCB ground via standoff screw. Grounding the pot shield plate to the black wire also helped.

Heater artificial ground connected to either a terminal strip chassis point at the AC cord or the above 'audio ground' doesn't seem to audibly different.

What is missing is the chassis ground connections Fender has drawn at the input and output jacks.

If the OT secondary low side should be chassis-grounded also, for safety reasons, I'll have to try that. I thought I did but don't remember.

As found and as-drawn, a chassis ground loop formed by two separate chassis ground connections several inches apart seems (in my ignorance) to be why the foil covering on the back cover exists. I have it pretty quiet without the cover on, using a separate speaker in the garage.

Back in the cab with its own speaker, and the cover installed soon. Hopefully no change in noise. Unless a reason for chassis ground at the jacks is convincing.

[CharvelBlue]

How did the amp sound when you got it back together? Less noise and hum? Lots of information in your post most of which I'm still digesting.

[Stevem]

Lets narrow things down please!

As for me when I see what I will term a long winded post and or when too many issues are listed / posted at once I am turned off from reading it.
No offense meant here!

So let's divide and conquer.

What do you want to tackle first ?

If it's the hum issue then first we need to know if it's 60 hz, or 120 hz.

This will then further narrow things down to being on the DC side of the power supply, or the AC side.

I would we do this first since the hiss issue is also on the D.C. Side of the supply and delt with last.

The ball's in your court!

[rooster]

I had one of these once, this year production. Yep, it hummed. I had received it from a Fender promo for free so I didn't have anything into it. After some typical experiments, thinking like the OP, that there was some layout problem. In 2002 there weren't aftermarket PTs or OTs made especially for the Pro Junior or Blues Jr. The store I worked in was a Fender dealer and we got a two more of these in: same problem. As a result we stopped ordering the Pro and discounted the price on the ones we had. I finally donated mine to the Goodwill.

So, years later I've revisited that hum and the particular amp (in mind only). I decided it was the PT. FF a couple of years and the Pro line was quiet. Did the build change that much? No. Just sayin.

[Stevem]

If the amp has a 60 hz hum out of the speaker when you first turn it on and even before the tubes are warmed up then it’s a case of the PT magnetic’s bleeding into the OT.


This issue is due to the PT being mounted too close to the OT, and or the Pt or OT not being oriented right in relation to each other.

You can prove this by removing both output tubes and the amp still hums at 60 hz.

All of the U45-B 1-12” model Jap made Univox amps have this issue and a few small late 50s and 60s American made amps have this issue also.

I forget if those Fender PTs in those amps even have the needed Copper shielding on them?

[CharvelBlue]

Yes they have the copper shielding. So it's safe to turn amp on without any power tubes?

PJ OEM output transformer 5.jpg

[Stevem]

Having copper shielding on the PT has nothing to do with the amp being safe to power up without the output tubes in it, the coppers effect is to block 60 hz magnetic flux from getting into the audio / gain sections of the amp.

In short, yes all is fine with powering up the amp without the output tubes in, but since the filiment voltage will then run high, it’s best to yank out the preamp tubes also.

[CharvelBlue]

Thanks for the info re pulling the tubes and yes I realise this has nothing to do with the shielding.

[Mant]

Well was it quieter?

[TUBEDUDE]

Having copper shielding on the PT has nothing to do with the amp being safe to power up without the output tubes in it, the coppers effect is to block 60 hz magnetic flux from getting into the audio / gain sections of the amp.

In short, yes all is fine with powering up the amp without the output tubes in, but since the filiment voltage will then run high, it’s best to yank out the preamp tubes also.

Yes, the xfmr copper covering is usually called a flux band.

[Yoda]

I’m curious how much quieter the amp was after all this work. I worked on a Pro Junior a couple times to reduce noise and still didn’t get good enough results so I’m simply rebuilding it with a completely different hand-wired circuit.

[Stevem]

Your doing in a way the same thing that the Op did in this string at the start, that being you want a detailed answer to your general question.

What noise are you having take place, Hum, buzz, hiss or some of all 3?

[maxkracht]

I rebuilt one of these recently, there's a post about it around here. It sounded like bad filter caps to me, but the caps tested fine. I didn't troubleshoot very far before deciding to start over. Easier to make something good from scratch... In my opinion, there is one too few power supply filter stages. I don't think that is THE problem, but it isn't helping. I can confirm that the noise isn't transformer related. My final product was quiet.