Issues with Printed Circuit Boards used in Guitar Amps

[SoulFetish]

> Normally removing sixteen screws would be a service no-no, but I put a removable panel behind the PCB in the chassis so a service tech could access the solder side without removing the PCB from the chassis.

all really good design ideas, but this one ^^^^ is the best (speaking from someone who services other people's amps and can't leave his own amp alone)
I can't tell you how often I wish I had a trap door access inside a chassis. I would even extend this to many turret/eyelet wired amps as well. This was especially true on my own amp when I thought it would make for a nice layout to do much of the wiring under the board. After all the careful layout I had drawn, and physical wiring was complete, I tracked down mistake I had made in the drawing where the connection was off by one turret (which coupled the next stage to the previous grid instead of plate). I was getting some strange voltage readings.... off by one friggen turret to the right. But what a pain in the ass to try and get to. If I make under board terminal to terminal wire connections it will not be done like this -

[Gregg]

When it comes to tube amp PCBs let's not forget these classics:

[Enzo]

While a poorly designed or super cheap PCB might be unreliable, I feel obligated to point out that the radio in my car is PCB based, and that thing fires up instantly when it is 100 degrees outside or when it is -40 degrees. Matters not if I drive rough roads. And those space satellites that went to JUpiter, Saturn Neptune, Uranus, and Pluto, all had circuits on PCBs, and out in space the temperature is close to absolute zero. Those Mars explorer thingies? PCBs inside.

[Zozobra]

This has been an interesting read as recently I have been messing around making my own turret layout style PCBs using the EasyEDA online editor. Their PCBs are pretty cheap so if I make something shitty it can go in the bin and I'm out $10. I've been taking a modular approach so the preamp is on one mini board and a tube buffered loop and LTPI is on another. I've been making efforts to keep cable runs as short as possible and mechanical layout sensible. Not everything has to be shoehorned on the PCB if it makes more sense to mount on the tube socket or the back of a pot. Preamp tubes are mounted at the front of the chassis a la SLO as this seems to be the best way to keep things tidy.

[Leo_Gnardo]

When it comes to tube amp PCBs let's not forget these classics

More secret messages can be found on PCB's from THD and Mesa. Sorry I have no photos available. But it is fun to find 'em.

[nsubulysses]

When it comes to tube amp PCBs let's not forget these classics:

[ATTACH=CONFIG]50512[/ATTACH]

[ATTACH=CONFIG]50513[/ATTACH]

I have a VHT pitbull built in 2001 that is built with MESA brand caps dating from 1992. High quality, top of the line stuff no doubt

I cannot get it to stop making intermittent crackling and whooshing/swirling noises, similar to a mesa. That's kind of a funny message right?

[SoulFetish]

When it comes to tube amp PCBs let's not forget these classics:

[ATTACH=CONFIG]50512[/ATTACH]

[ATTACH=CONFIG]50513[/ATTACH]

That is awesome

[Chuck H]

Well then I wouldn't bust Steve Fryette's gnards for not using ceramic tube sockets or mounting power tube sockets on the PCB, but I might ask him about using nine year old filter caps in a new amp. I think those old Mesa caps are Sprague. Basically custom Atom caps. And I was told that back in '92 Sprague had an expiration date assigned to their batch codes. Rotation of stock was requested and returning product not sold by a specified date. I don't know how long from manufacture that date was, but I'll wager it wasn't ten years. Those caps in that amp are twenty six years old. And being unused for the first nine would actually be more detrimental than if they were in an amp getting some use.

[Gregg]

Hardly any cap would start getting used soon after they leave the factory. Depending on the cap several years is not such a big problem I think. AFAIK these cap are cheap low quality anyway.

[Chuck H]

I've installed decade old electrolytics in a build and they exhibited some of the common failure symptoms almost right away. And it got a little better as the caps were used more, but never completely so I replaced them with new ones. I was also present for the unboxing of a twenty five year old amp that was still in the original manufacturers packaging. I warned against just firing it up because of the filter caps age. The guy did it anyway and it took about five minutes for the caps to explode. Nine years, or even several years is too long. If you build a lot of amps you can keep inventory in rotation and you build few then you probably don't have too much capitol tied up in inventory, so you buy fresh for each build. That's my position because personal experience has shown me that deviating causes problems.

[Leo_Gnardo]

I've installed decade old electrolytics in a build and they exhibited some of the common failure symptoms almost right away.

What if you put years-old caps thru a forming session prior to installation?

I have a box full of 400 uF 450V "beer can" caps dated 1966. Anybody want to set off some big fireworks? Should be fun. As it says on the old fireworks packages "stand well back."

[R.G.]

Cap shelf life, old caps and reforming caps:

Most cap makers specify a shelf life. The last one I remember was about five years.

The electrochemistry to grow the oxide layer used to be the same as the one in the cap, back when cap electrolytes were aqueous. I believe that today, the oxide insulation is grown in a separate and somewhat different forming bath than the electrolyte that fills the caps. This most likely means that re-forming is not as good as it once was when the electrolyte was more like the forming bath. That line of reasoning leads me to believe (based on the thinnest of evidence) that re-forming is not as good as it once was.

There is a re-forming process that gives you the best chance of success. Re-forming depends on letting electricity leak ...gently... through the pinholes and weak spots so the current pulls in the chemical reactions that form oxide. It was once common to put a high voltage power supply equal to the voltage you wanted to form to across a 100K (or so) resistor and the cap you're trying to form. A voltmeter across the resistor or cap tells you when the cap is up to withstanding 90% percent of the forming voltage, and you're fairly safe assuming it's formed enough for the full forming voltage not to damage it.

Damage happens when the current density in conducting spots in the oxide heat up that spot so much that it burns through the spot, instead of growing oxide. Keeping the current density down (that 100K resistor) was important to getting re-forming inside the cap.

It worked the few times I tried it, excepting for a few notable failures. Electrolyte smells bad.

Seriously, caps are so much better, put in new ones.

[SoulFetish]

Seriously, caps are so much better, put in new ones.

Well, this certainly should be the case. Manufacturing process and materials technology has advanced significantly. Yet, I've tested examples of electrolytics pulled from vintage amplifiers which (much to my surprise) have maintained tight tolerances and lower ESR measurements than new production caps. In the case of the modern axial Illinois electrolytic caps, I've repaired many failures which occurred in a fraction of the time you would expect to see.
I admit, this would not qualify as any kind of scientific testing, as no data was recorded or test sample being large enough. It was more for my own edification. Plus, I would add the caveat shared by someone wiser than I - "As with all anecdotes, it reinforces my biases."

[Chuck H]

Yet, I've tested examples of electrolytics pulled from vintage amplifiers which (much to my surprise) have maintained tight tolerances and lower ESR measurements than new production caps.

Based on what R.G. said, that the electrolyte in the old caps WAS the actual solution used to grow the oxide layer, it seems likely that caps using this older production method would tend to repair themselves for years and years as long as the chemicals hold out and the amp is used semi regularly. It would certainly explain why there are 50yo caps in amps that are still working. I know with certainty that I wouldn't expect that from any standard line manufactured today. I was at a loss to explain it but now I think I understand why

[R.G.]

Well, this certainly should be the case. Manufacturing process and materials technology has advanced significantly. Yet, I've tested examples of electrolytics pulled from vintage amplifiers which (much to my surprise) have maintained tight tolerances and lower ESR measurements than new production caps. In the case of the modern axial Illinois electrolytic caps, I've repaired many failures which occurred in a fraction of the time you would expect to see.
I admit, this would not qualify as any kind of scientific testing, as no data was recorded or test sample being large enough. It was more for my own edification. Plus, I would add the caveat shared by someone wiser than I - "As with all anecdotes, it reinforces my biases."

Once upon a time, I was extruded through a course in engineering statistics. One of the things they beat into us, along with the bathtub curve, was the idea of survivors' bias. The idea was that you can't necessarily trust the few old [whatevers] you look at as representative of the class of things because the "weak" ones have already dropped out. A more subtle consequence of that selection process is that the survivors may not tell you much about what the whole class looked like originally. By their very nature, the survivors are not like the ones that already died.

As for the modern ones failing, sure - you're seeing the infant morality side of the bathtub, all the early failures of the weak ones, the survivors of this generation simply working and effectively hiding in place, doing their jobs.

It was common back in the 60s to have manufacturers list the tolerance on bigger electrolytic caps as -20%, +80%. A specified tolerance that wide tells you that they wanted to sell all of the caps they could make, even with crude control on the resulting distribution of parts, and would just guarantee you "enough", not accuracy.

A significant part of the work on modern electro caps was put into making the caps smaller for a given capacitance and voltage. That meant a push for thinner foils, etching the foils for greater surface area, and for less electrolyte and more effectively electrolytes. Thinner foils tended to push ESR up, but by changing the winding methods to extend the foils along the winding, the caps could have more taps on the winding and reduce the ESR again. Also, better manufacturing controls meant that the wound foils/separator was more predictable, and the makers didn't need to leave as much spare room in the can for variances in the winding. The caps got tighter.

Humans are going to find a pattern in any data. I know I do.