Fender Hotrod Deluxe Parasitic Oscillation

[Valvepower]

Hello,

Just seen this post and if all else fails this could be worth a try.....

I remember many, many moons ago a similar problem on a similar kind of phase splitter. This was solved by placing a 4700pF and 270R in series and then putting this series network in parallel to the 4K7 (R68 in your circuit...?) which goes to ground.

Could increase the 4700pF to 10nF and the 270R to 470R...?

This is of course assuming the presence is at minimum (0) when the potentiometer is at its maximum resistance! I.e. decreasing the resistance increases the presence.

Regards
Terry

[kevinaston1]

Just seen your post.

A few years back, a friend of mine rung me relating exactly the same fault. Without thinking much, I said if you have oscillation, you have positive feedback. He gave out a strangled groan and said, yep, someone has changed the output transformer recently.

Well I said, reverse the feedback winding. He did, and the osillation was gone.


Kevin

[frankmcvey]

Hi, All,

Well, life got in the way, as sometimes happens, so sorry for my delayed response.

This morning, I tried Terry Valvepower's suggestion of making a series network of 4700pF and 270R, then placing that network in parallel with R68 (see cct diagram in post#1), and that completely cured the parasitic oscillation problem.

At the moment it's all croc-clipped together. I'll get my mate in with his guitar and we'll check that the modification doesn't affect the sound too much. If all is OK I'll figure out rather more elegant way of incorporating the mod.

Any theories as to the cause of the problem, and why that should have solved it? Is it all down to just slight phase shifts?

Thanks everyone for all your suggestions, and a nice cup of tea and a sticky bun to Terry, who solved it (or at least gave me a way to paper over it!)

Cheers,

Frank

[Herald1360]

Originally Posted by frankmcvey

Any theories as to the cause of the problem, and why that should have solved it? Is it all down to just slight phase shifts?

Cheers,

Frank

An amp that bursts into oscillation on signal peaks is likely to be in a state known as conditionally stable. One of the quirks of feedback and oscillation is that the condition for oscillation is unity gain round the loop when the fed back signal is in phase with the forward signal at the feedback injection point. Both these conditions must be met for oscillation. So, somewhat counterintuitively, if the loop gain is greater than unity as the phase goes through zero the system is still stable and the whole system will be stable provided that the phase shift rises again through zero before the gain eventually falls to unity. Wonderful!


Well, no, not really. If the amplifier is driven towards clipping, at the commencement of clipping, the effective gain falls and you hit a point where the oscillation criteria are met...... Squawk! Since clipping is unlikely to be absolutely symmetrical, the amp hoots on one peak but not the other.


The network you added has probably either modified the loop response to roll it off before the phase crossing point or if it's done well, it may have added some phase advance without too much gain increase and will defer the phase crossover to a higher frequency than the point at which the gain crosses over.


And yes it is down to phase shifts which may be quite incrementally quite small, just at precisely the wrong place in the frequency spectrum. It's possible that component value drifts elsewhere in the amp have resulted in the problem if it hasn't always been there.

[McMurdo]

I've got one of these on the bench at the moment for an unrelated fault. I'll see what it looks like on the scope. This model is a world-wide seller and any issue with design would be well documented, I'm sure.

Dry joints and microphony are the usual suspects of instability with any 'combo'.

[theoldtrout]

These amps are notorious for instability. Fender use some poor quality electrolytics in these amps, I found changing the HT line de-coupling caps and dressing those horrible ribbon cables sorts them. Also microphonic valves make them oscillate.

[frankmcvey]

Well, after all that, the owner wasn't too bothered about the PO, since the onset was above his usual gigging volumes and the effect on the sound seemed to be pretty minimal in this case. As Chris said, it may well have been like that for a long time.

Chris Herald - thanks for the lucid explanation re the parasitic oscillation problem. To be fair, you did mention it your first post!

Kevin - yes, there are a few design faults with this model, and they're fairly well-documented on the internet. TheOldTrout refers to a couple of them in his post.

FWIW, I've attached a copy of my report which collects together most of the stock problems with this model and may be of some use to others in the future. It's here. The very last page is the notice I stuck inside the amp for future techs digging in to it.

Thanks once again to all for their helpful input.

Cheers,

Frank

[McMurdo]

To be fair dry joints around hot components or components used for mechanical anchorage, control pots etc are pretty normal for any piece of equipment, not specific to the Fender. This version I have here has rather good 'IC' branded capacitors (Illinois Capacitor). Like any combo, the valves particularly the first pre-amp can go microphonic or require a new one being 'selected' for low noise...gone are the days of Mullard quality and consistency.

The two faults on this example were intermittent thermal runaway caused by one of the bias voltage reservoir caps having a dry joint, and the other was a faint rustling caused by one of the l.v rail smoothers having a hairline crack in its legs; being held next to a neighbouring component only by hotmelt glue.
Both caused by vibration-induced fatigue no doubt.

No tendency to oscillate on the scope right up to the onset of o/p stage clipping

[Valvepower]

Hi,

Just returned to this thread..... As Frank eloquently said life got in the way, so sorry for the delayed response.

That’s good it solved the problem, but I can understand not implementing it as the user may never encounter it the way they use the amplifier. I’ve had situations where I’ve found problems like this, which I encounter on the bench, but the user hasn’t encountered or commented on under normal use.

I feel Herald1360 hit the nail on the head in his last paragraph and I’m wondering if it’s always been there and has developed due to component spread or just wasn’t found on test as it would have been run into a resistive load. I figure the amplifier would have been given listen test, but not at the level required to get it to oscillate.

I treated myself and SWMBO to a piece of Victoria sponge and a coffee at the local RHS gardens

Regards
Terry