DRRI Power Supply Capacitors

[Chuck H]

Well...

I don't build often enough to volunteer for this, but I think a test is in order. A comparison of a pair of 68uf caps in series vs. a 47uf and a 220uf in series. For the test we would want the balancing resistors to have the same net value, I think. So a pair of 270k or 300k resistors for balancing the 68uf caps. Test for ripple at that node. If possible test for discharge and recovery time under normal use conditions. Look at the output wave form at under heavy use and compare, etc. And of course the play test to see if there is any difference in feel or tone. For the purposes of the test I think all caps should be of the same model/manufacturer. And, of course they should be installed in the same amp with a switch (careful! YIKES).

[Pedro Vecino]

FWIW the Engl "Straight" 100W amp head early 90's also had dissimilar filter cap stack much like the one being discussed.

Engl Straight wherein I have uses old blue double axial F&T electrolytics. Stacked and symmetric in the first two nodes.
"Modern" Fender (Pro Junior, Blues Junior, HR Deluxe, HR Deville, Blues Deluxe and Blues Deville, Supersonic) and reissues (Deluxe Reverb, Super Reverb, Twin Reverb, 65 Reverb and Bassman) only uses high voltage electrolytics (above 300V ) of 22, 47, 100 and 220uF (this last only in Twin Reverb). 33 and 68uF are standard values anywhere but to my knowledge in the last twenty years Fender does not use them. Or at least I have not seen.
Adapting electrolytics of dissimilar value in series with paralleled inversely proportional resistors does not seem very orthodox but electrically works well.
Applications of this type are seen at times. Laney uses in many models two 4n7 capacitors in series to get something close to 2n2.

[Enzo]

Test away. Something you might test for might have no bearing on their reasons. What if it turns out to be some rectifier life thing, how would your test reveal that? Or maybe it stopped some occasional fart noises under odd conditions? I think it hard to test after the fact.

I propose another test, I am going to call someone at Fender and ask them.

[Enzo]

I just spoke to my pal at Fender service. I don't usually like to bother the factory guys, but...

His basic response was "I have no idea." He thought maybe they found it more stable under some conditions, but mainly he told me that such arcane engineering secrets don't always filter down to the shop level.

Sorry. I had to hang up early, as I have an incoming call. I think he was about to tell me that all MY ideas were good ones, and everyone else was probably wrong. But I am not sure of that...

[g1]

I emailed the question to them yesterday, will be interesting to see if I get the same response. I did include the word "design" in the subject line, in hopes of getting to an engineering type.
However, what I'm expecting to hear is "Were you talking to Enzo? He hung up on me just when I was about to tell him he was right" .

[Chuck H]

"Modern" Fender (Pro Junior, Blues Junior, HR Deluxe, HR Deville, Blues Deluxe and Blues Deville, Supersonic) and reissues (Deluxe Reverb, Super Reverb, Twin Reverb, 65 Reverb and Bassman) only uses high voltage electrolytics (above 300V ) of 22, 47, 100 and 220uF (this last only in Twin Reverb). 33 and 68uF are standard values anywhere but to my knowledge in the last twenty years Fender does not use them.

This is the most significant support for this design being a cost consideration while achieving a desired uf value. It's entirely possible that Fender gets a price break based on specific values. Having worked in both retail and wholesale I saw this all the time. A customer would require an extra special discount on one or a few particular items and their sales rep would arrange it. But, because of the obscene price break on the selected items, if that same customer had to buy something other than those items they paid the normal price. To get the extra special price they would have to place an order of similar magnitude for an item as well as the assurance of continued orders.

So it's a price break issue.

[R.G.]

It could be any number of things. I have a friend that worked in the RCA broadcast division. Back when there was a broadcast division. Back when there was an RCA that was more than just a brand name. He could look at any schematic in their shop and tell who designed the circuits. This was because there are always some components that don't have any overwhelming need to be a specific value. Some resistors just need to be a resistor, and where one would use a 1K, anything between maybe 680R and 1.4K would work OK. The guys in the RCA design shop used resistor values as signatures. One guy would always use a 1.1K, 11K, or 110K for unimportant values. Someone else would use 910, 9.1K, or 91K. This was apparently informal, but strongly observed.

[Enzo]

No trouble believing that.


I ran a shop in the 1970s with seven or eight technicians in field service. I could recognize their work. I knew each guys solder style and how he dealt with wire ends, and other little tells.

Whether purposeful or not, we all leave our signatures in many places.


We have a bunch of surplus stuff here, got from the MSU world. Thinks like precision resistors. 102k or 1631 ohms. A guy we used to work with was real anal about parts. He'd buy the same 4558 ICs from PV, from Fender, from Crate, from Yamaha. He was convinced all those places carefully selected THEIR 4558s to work in THEIR circuits. Later, when we acquired all these odd parts, my partner Bill and I planned this amp design. Nothing special, Bassman sort of thing. But it would use those 102k resistors in the 12AX7 plate and the 1631 ohm as cathode resistors. We wanted to arrange someone wind up taking the amp to the anal guy tp drive him nuts trying to figure out why all those parts needed to be the off values. Alas, we never got around to it. I suspect it was funnier in our minds than the reality would be. Besides, we wouldn;t have been there to see the result. But I know for sure, he'd have been looking for sources on those values.

[booj]

It's cheaper. Low voltage caps are cheaper than high voltage ones. so you can use one cheap cap and one expensive high voltage cap and get away with it.

[tedmich]

Hee Hee This is like when a group of smart hipsters with advanced degrees get together and seriously discuss arcane plot points from hastily put together crap dramas like "Lost" (or interpret Bible verses) Humans make dumb ass mistakes constantly, over analysis is not fruitful. You want complexity that WORKS look to nature; human complexity is usually just a silly mistake, that MAY work. For both human and natural endeavors if it doesn't have an obvious negative effect it often persists just out of laziness. Maybe this cap arrangement brings one closer to God and/or or sounds awesome...

[bob p]

I think everyone is over thinking this. You're looking for an engineering-based reason for what appears to have been a cost-containment measure. I don't think you'll ever find an engineering-based explanation for such a screwy totem pole. In my opinion the answer is simple -- some bean-counter figured out that they could lower the amplifier's cost by constructing an oddball totem poll that sourced the cheapest combination of commonly available parts that would do the job.

To check on this theory, I showed the schematic to a 90 year old friend of mine who used to be a design engineer at Motorola during the Golden Age. I remembered him always griping about how his team in the R&D department would strive to design the best circuit possible, and their best designs never made it to production. Why? Because when the R&D team finalized the design, it would be passed off to a production team whose job it was to squeeze every possible penny of cost out of the circuit's design. His opinion was that this type of totem pole setup would never have been chosen by the Development team; it served no design purpose and the implementation was too obtuse. In his opinion the circuit change was obviously a cost-containment measure that was implemented by the Production team.

At Motorola the Production team would then tear the circuit apart, finding every possible method of cost reduction. This invariably involved compromises to the circuit's performance in favor of signifcant cost reductions on a production run. (Think of how GM struggles to shave 5-cents off of the cost of producing a component for a car, and how that 5-cent savings gets multiplied by millions of units.) It was not at all uncommon for the Production guys to make significant changes to a circuit to shave off production cost, sometimes even trading away a circuit's best performance features to render a less expensive product with "good enough" performance. The result was that the R&D guys were always mad at the Production guys for screwing up their designs.

In this context I think it's pretty obvious that Fender decided to use this totem pole arrangement so they could buy the cheapest commonly available parts that would fit the circuit's performance needs. there's no other reason to choose such oddball values.