Modulus "Deepfried Marshbar" build debugging!

[JMPGuitars]

You can try swapping pin 7 on the power tubes, but that might make it worse. It doesn't usually matted without NFB, but you have a NFB so it does matter.

Follow the modern ground scheme thread linked in my signature.

Also post your voltages. As I mentioned previously, your power tube cathode bias resistor should probably be 180Ω.

You may also need to do some chopsticking.

Thanks,
Josh

[Iain T]

Hey Josh,

Thanks again for the reply!

That grounding scheme is pretty much what I was considering - in the schematic it's be as simple as cutting the under-board ground wire that goes to the 22uF cap in the pre circuit and moving the ground point from the middle pot over (or routing under-board to the right instead of left.

Voltages for V4/5
2: 9mV / 12mV
3: 11.72V
7: 328.2V / 326.3V
9: 319.2V / 320.2V

If I've got this right the tubes are biased pretty hot at 11.7W dissipation, 97.5%.

Then again, I might be close to 100% wrong as I'm still trying to work out how to calculate tube biasing!

[JMPGuitars]

You calculated correctly. I assume you used this: https://robrobinette.com/Tube_Bias_Calculator.htm

Anyway, yeah, replace the cathode bias resistor with a 5W or 10W 180Ω. To be safe, you might want to get a couple higher values too, like 200 and 220Ω.

You might also want to consider increasing the screen-grid stopper resistors to 1k.

[Iain T]

You calculated correctly. I assume you used this: https://robrobinette.com/Tube_Bias_Calculator.htm

Yeah, that site is an absolute gold mine of info. through most of it before even thinking of building an amp.

I'll have a look at the other bias and grid-stopper changes once I've tried swapping pin 7 over then review the grounding scheme.

Cheers!

[Iain T]

Pin 7's swapped... immediate squeal so I did have that the right way round. My guess is that they updated their output transformers (Modulus wind their own) and colour-coded the two to brown and blue as per the layout. You can buy the kit without the xformers so they are probably catering for ones that are just red-red for the two ends of the winding.

Re-wired the output trans and re-tested. Measures clean outputs at the speaker terminal into my dummy load (measures 4.45 ohms) and got the following peak-peak voltages with a clean sine wave:

Low: 28.8V (p2p)
High: 28.4v (p2p)

This comes out, as I've muddled through the maths at ~23W, a touch high for this alleged 20W amp. Increasing the bias resistor definitely makes sense!

Re-checked the voltages as well and have, V4/V5 and this confirms it's running way too hot... I assume, being able to hear the sine wave in the amp under full load into the dummy isn't ideal and is a sign of the amp running too hot?

2: 3-5mV
3: 11.85 idle, 13.67V under load. That's 99.2% / 113%!!!
7: 314V
9: V4: 322.6V / V5: 324.4V

Other than the above it's sounding great now, I retouched some solder joints and this has improved things a bit. Will adjust the bias and review the grounding scheme but I'm currently very very happy.

[JMPGuitars]

Your target is 85% at idle voltage when biasing the power tubes.

I figured it wasn't the plate connections (pin 7) because you said it didn't squeal all the time. Now you know for sure. Brown/red/blue are common for 18W OT primaries.

What you described sounded like a parasitic oscillation, but there's little point in too much troubleshooting that with your power tubes running as hot as they are, as that could be part of the problem. Your ground scheme can also be a big part of that problem.

Once you have the power tubes tamed, and the ground scheme corrected, if there's any angry noise left, you should hopefully be able to shift some wires with a chopstick to eliminate it.

[Iain T]

Thanks - I'm fairly sure the lead dress is not optimal, it's a fairly tight space to run everything in and I opted to keep the leads as short as possible, a bit more length might actually allow for cleaner routing.

I have some 10W cement resistors on order - 180R, 200R and 220R as well as some ground lugs to work on the grounding scheme.

I'll check back in in a week or so when this is done.

Again, thanks for taking the time to share your experience, it's appreciated.

[Iain T]

Long weekend away so not had any time to work on this until tonight!

Just run through 180R, 200R, 240R, 220R in that order until I landed on a bias that works...

13.65V with the 220r - 87.5% and that seems to have got rid of noise on the power stage completely. Gan on 0 and master up full and it's really quiet, almost silent.

Add any gain and it gets noisy quickly - hiss and interference so grounding scheme and lead dress a must to improve that.

Would adding a 470pf cap to ground off the 10k input resistors help with the hiss? Not quite sure how I'd route it with this layout but must be possible!

[JMPGuitars]

I don't think you shared any photos of your build. Can you post some? What type of resistors did you use?

Thanks,
Josh

[Iain T]

Hi Josh,

The Bias resistor is a standard 10W cement resistor. The 10ks on the inputs are 3W Metal Oxides.

Will do some more work on this on Sunday and take some detailed pics, here's one from before changing the bias resistor - right-click and open in a new tab for higher res view! Anything you'd want a close-up on?

Thinking of the following ground points:
1: Safety ground lug (as it is)
2: Centre-tap ground lug near the safety.
3: Power-stage ground lug near the left front.
4: Current input/pre ground lug.

2 and 3 could be a single ground point though, not sure if there's much to gain or lose with either.

[JMPGuitars]

I would replace your input resistors with CMF metal film resistors. That includes the 1M, and the grid stopper resistors. Carbon Film isn't as noisy as Carbon Comp, but hiss is usually related to resistor type choices. Metal film being thing least hissy of those choices.

See the thread in my signature for ground scheme improvements.

Thanks,
Josh

[Iain T]

I would replace your input resistors with CMF metal film resistors. That includes the 1M, and the grid stopper resistors. Carbon Film isn't as noisy as Carbon Comp, but hiss is usually related to resistor type choices. Metal film being thing least hissy of those choices.

See the thread in my signature for ground scheme improvements.

Thanks,
Josh

Thanks Josh - already digested the grounding thread. Came across an interesting variant from Lyle (Psionic Audio) where he likes to ground the centre-taps on certain Marshalls separately from the power and filter stage ground. Going with your suggested route is by far the simplest for me as it only requires breaking one link on the existing ground bus and a single hole for the lug!

I'll look at getting some alternate resistors ordered.

It'll be interesting going through each step of changes to see what yields biggest gains!

[Iain T]

Initial attempt to improve the grounding scheme was sub-optimal... much louder earth hum!

Having checked against the "18W Lite 2b" schematic I realise that I had the PI grounded with the power stage via the presence/resonance controls and I hadn't properly traced it through. Swapped that to ground with the pre-stage and it was back to around the same level of hum as per original scheme.

One change - much more noticeable interference that was affected by moving my hand over the input area while the chassis is out of the head cabinet. This wasn't noticeable with the original ground scheme of everything grounding by the input...

Did a full volume test yesterday, pre-changing the grounding, at a local rehearsal space into a 2 x 12 with Celestion V30s... My guitarist friend was seriously impressed by the tonal options and sheer volume.

Amp was much quieter there than testing on my bench at home... realised that the powerline adapters I use for getting internet up to my workspace were adding a huge amount of noise to the mains!!!

Anyway - with master and gain on full using the tone switch in "#34" or "800" positions, there is a lot of hiss. Waiting on the metal film resistors to see how that is impacted.

[Bieworm]

Hi.. I think hiss is inevitable with this circuit. Ever heard a real plexi when dimed?

[JMPGuitars]

Hi.. I think hiss is inevitable with this circuit. Ever heard a real plexi when dimed?

You're correct, however, not using metal film resistors at the input stage will add hiss to the circuit.

CF resistors have less hiss than CC resistors, but more hiss than MF resistors. Any hiss at the input stage gets amplified throughout the circuit.

[Iain T]

Hi.. I think hiss is inevitable with this circuit. Ever heard a real plexi when dimed?

Ha, very very good point, maybe my goal/expectations were too high!!!

I've swapped the 1M and 10k resistors on the inputs for MF. I also took a minute to re-route the low-gain shielded input so it crossed fewer other wires going to V1 - they just cross their respective cathodes with as much vertical separation as possible around 1cm.

I also took the time to completely re-do the input jack wiring. I wasn't happy with the initial job I'd done. It was all sound but messy.

Anyway - master dimed and there is just a little hum.
With the gain and master dimed - low-gain input is just a little hiss, more on the Deepfried/#34 setting, high-gain is more-so but understandable.

I'm not sure that the swap to MF resistors has made much difference but a worthwhile exercise.

On the high-gain channel, with the gain at full and either the DF or #34 circuits engaged it's too much gain for even a good thrash sound but, surprisingly, there's still decent definition of the sound - it doesn't go to flub or fuzz which is a surprise.

Next job it to measure the output on the 'scope and see what clean power we're putting out.

Then to work out what to build next - probably a simpler version of this and on my own, from scratch, rather than a kit-build.

Thanks so much for the input on this guys - it's been invaluable!

[guitarmike2107]

Hey Iain, Your build looks great well done.

The design of this kit is heavily based off the BE mod, that is why the gain control is after the second gain stage.

The grounding is already optimised as each part of the circuit is connected directly back to the part of the power supply powering it with no shared ground paths. Grounding at different points on the chassis is not normally recommended these days as it forces current to flow in your chassis and makes the ground scheme quality dependent on the quality of the chassis connection. Ideally your chassis should just be a shield.

The hiss noise is largely the result of the large 330k plate resistor and the 10k grid input resistor. There is a reason Trainwreck and Mesa did not fit the grid input resistors. Changing them to metal film does not make much difference. Reducing them does though, so if the amp has more gain than you need you can drop the first gain stage plate resistor down towards 100k for less gain and less hiss. Changing the plate resistor does also change the compression/feel a bit.

The presence pot scratches just like it did in 1965, but has a smoother effect on the presence for the full range of the pot, I prefer it which is why I used it in the kit. You can change it to the version used on the 800 series amps, but then the pot does little for a large part of the travel.

Cathode biased amps can run at 100% plate dissipation happily, When you play them they actually drift cooler, the opposite of fixed bias amps. Of course if you are happy running your valves cooler then that's great. The B+ voltage in these amps is lower than the original 18w amps which often needed a larger bias resistor. Just watch your wall voltage, It can be anything between 230v and 248v in the UK and that will affect your bias.

If you have oscillation changing the pink wires on the gain and master volume control to shielded wire may help, Also upping the 47pf capacitor at the phase invertor can help, You can go up to 250pf without it affecting the tone.

Hearing the amp when plugged into a dummy load is normal, Is is due to magnostriction. A quick google search will explain it well.

If you play metal and thrash have a look at the SAT switch/Jose Master mod which adds clipping diodes. More gain and compression with a slightly different feel.

If you want clean you need to plug into the low gain input and use the 800 position for the lowest gain. You should be able to find some nice clean sounds that way.

I learned loads on this forum almost 20 years ago, There is a wealth of information here. loads, try mods, but one of the best ways to learn is to help others fault find there amp builds. You start to spot mistakes and errors easily and learn off others mistakes.

Thanks, Michael

[JMPGuitars]

The hiss noise ... grid input resistors. Changing them to metal film does not make much difference.

That's true if the material choice isn't influencing the hiss problem. If the amp is all carbon comps, it will absolutely hiss more than an amp that's all metal film.

If the input resistor material is a source of hiss, then changing it will reduce the hiss. Those input resistors get amplified through each stage; that's why the choice of resistors there is important.

[Iain T]

Hey Iain, Your build looks great well done.

Thanks - the instructions and photos on your site are really helpful and the neatness made it feel wrong to just throw the kit together! A few things I'll do differently next time to make life easier.

Thoughts on solid core wiring vs stranded? I kinda like the look and ability to route wires accurately with solid core. The heaters are the bit I'm most dissatisfied with - started twisted and then went to parallel as it was easier to manage!

The design of this kit is heavily based off the BE mod, that is why the gain control is after the second gain stage.

The grounding is already optimised as each part of the circuit is connected directly back to the part of the power supply powering it with no shared ground paths. Grounding at different points on the chassis is not normally recommended these days as it forces current to flow in your chassis and makes the ground scheme quality dependent on the quality of the chassis connection. Ideally your chassis should just be a shield.

Changing to splitting earth between the power and pre stages made no difference to noise levels - certainly no noticeable improvement.

Any thoughts on the benefit of lifting the heater centre-tap and creating an artificial CT?

The hiss noise is largely the result of the large 330k plate resistor and the 10k grid input resistor. There is a reason Trainwreck and Mesa did not fit the grid input resistors. Changing them to metal film does not make much difference. Reducing them does though, so if the amp has more gain than you need you can drop the first gain stage plate resistor down towards 100k for less gain and less hiss. Changing the plate resistor does also change the compression/feel a bit.

Lyle (Psionic Audio) mentioned putting a small capacitor between the grid and ground to reduce hiss - it was mentioned in on of his videos in passing so no specific value, is this an option in this circuit?

The presence pot scratches just like it did in 1965, but has a smoother effect on the presence for the full range of the pot, I prefer it which is why I used it in the kit. You can change it to the version used on the 800 series amps, but then the pot does little for a large part of the travel.

The way the presence and resonance works on this amp is perfect. Why more amps don't have resonance is beyond me!

Cathode biased amps can run at 100% plate dissipation happily, When you play them they actually drift cooler, the opposite of fixed bias amps. Of course if you are happy running your valves cooler then that's great. The B+ voltage in these amps is lower than the original 18w amps which often needed a larger bias resistor. Just watch your wall voltage, It can be anything between 230v and 248v in the UK and that will affect your bias.

This will be something I'll experiment with!

If you have oscillation changing the pink wires on the gain and master volume control to shielded wire may help, Also upping the 47pf capacitor at the phase invertor can help, You can go up to 250pf without it affecting the tone.

I was wondering if this would be a smart idea anyway. I'd like to see just how hard this can be pushed without becoming unstable or excessively noisy.

Hearing the amp when plugged into a dummy load is normal, Is is due to magnostriction. A quick google search will explain it well.

If you play metal and thrash have a look at the SAT switch/Jose Master mod which adds clipping diodes. More gain and compression with a slightly different feel.

If you want clean you need to plug into the low gain input and use the 800 position for the lowest gain. You should be able to find some nice clean sounds that way.

Funnily enough I was considering this mod and looking at it in conjunction with the way the current 800/#34/DF switch works for even more flexibility. I'm assuming the existing switch on this amp is a take on the "cold clipper" mod that's I've seen on Rob Robinette's site with an even more aggressive set of values?

The tonal variations on this amp are pretty staggering - I'd be tempted to re-house it in a bigger chassis with some fun foot-switching options to effectively give a normal/lead channel switch for live use or dual FS to switch the tone options.

I learned loads on this forum almost 20 years ago, There is a wealth of information here. loads, try mods, but one of the best ways to learn is to help others fault find there amp builds. You start to spot mistakes and errors easily and learn off others mistakes.

Thanks, Michael

Really appreciate the feedback and info - forums like this are an absolute gold-mine of info and the people here are really helpful.

Thanks,

Iain

[Iain T]

If the input resistor material is a source of hiss, then changing it will reduce the hiss. Those input resistors get amplified through each stage; that's why the choice of resistors there is important.

After making this change I couldn't really tell any difference. It did give me a good excuse to tidy up the input jack wiring though!