Changing caps to get a more transparent, sparkly, rich CLEAN

[erwin_ve]

that the 1st would be the stiffest followed by the others. maybe 3 and then 2?

However, I'd think that if you got caps from the same series and took a b+ that was safe for all of them and ran a 500vdc cap, 2 250vdc in series and 4 250vdcs in series parallel.

the series parallel would be the stiffest followed by the 500vdc followed by the 250vdc.

So basically you're saying the same as me? Was there a misunderstanding regarding the precision word?

[diagrammatiks]

It's hardly the technical definition of precision. It seems that makes it a misnomer.

I never said it couldn't be stiffer then a single cap supply. I'd still argue that if you made it a strictly comparable situation the case would be different.

The issue is that in terms of high voltage supplies it's not always possible to switch between comparable equivalents.

If the b+ is 500vdc then a single 500vdc cap is going to be stressed pretty hard while 2 385s aren't. 2 500vdcs would behave even better in place of a single 500vdc. In that situation a 500vdc cap is just a bad idea.

If it was 10 caps in series parallel that would fit.

Fuchs has a great interview about the supply in the train45.

[Colossal]

Fuchs has a great interview about the supply in the train45.

diagrammatiks, do you have a link to this article? I'd be interested in reading it.

Thanks in advance.

[Bob Simpson]

Fuchs has a great interview about the supply in the train45.

diagrammatiks, do you have a link to this article? I'd be interested in reading it.

Thanks in advance.

I also am interested.

TIA

Bob

[diagrammatiks]

Oh man maybe andy will come in and remember where he talked about it...

I'll keep trying to find the google search combo that takes me to it.

I'm really hoping what I'm remembering is an article and not somewhere in the middle of a tgp thread.

This might take a while TGP only lets me see the last 250 posts and andy's got over 5000 over there.

I also remember reading an interview with Maven Peal. That was one of the first amps I remember seeing with Solens.

[img:210:180]http://www.mavenpeal.com/images/photos/zeetaheart.jpg[/img]

[llemtt]

@diagrammatiks

I'm not too much into "clean sounding" amps, but I somehow liked a Carr Rambler I once played

can you explain to me the technical reasons why a power supply filter (plastic?) film capacitor is better to obtain a "clean sparkling sound" than an "equivalent spec" electrolytic capacitor?

thanks
teo

[erwin_ve]

Well this is what Andy fuchs is saying on his website about the Ps filtering on the train 45:

Audiophile style circuit techniques bring out the best in any tube circuit. Guitar amps are no different. The Train-45™™ starts with an audiophile style high voltage storage system. Instead of using single large capacitors, the Train-45™ has a group of smaller capacitors in a bank. By using a bank of these smaller capacitors ganged together to produce one large capacitance, you end up with a power supply that is both forceful yet agile. One that holds lots of energy for transients yet can respond quickly to each and every note.

[diagrammatiks]

I think that at this point I'd be a little cautious about calling it a clean sparkling sound. There's plenty of amps that can fit that description. More importantly, a film supply isn't going to make an amp sound like an svt. Big changes in the circuit are always going to take precedence over component changes.

I think it's easier to think in terms of what contributes to the carr sound or what should happen if the power supply in the carr was changed and why.

In terms of measurable characteristics the ESL and ESR of a film cap is much lower any electrolytic capacitor. I haven't seen published numbers for Solen but Mundorf lists 9mohm ESR for their 100uf polyprop cap. This translates to a lower leakage and lower impedance power supply overall.

An electrolytic capacitor has on average around 250-400mohm ESR for a 100uf cap of the same voltage and while there are wide swings in the specs of different series of electros none of them approach single digits in comparable values.

If you google esr/esl there are a lot of articles and white papers that can explain the technical aspects. In particular, articles about igbt snubbers and the transition to film for dc-link capacitors.

avx corp has as good set of whitepapers

http://www.avx.com/techinfo_doclisting. ... Inductance

kemet has a great set too

http://www.kemet.com/kemet/web/homepage ... l%20papers

There are a number of industries that are actively trying to move away from electrolytics and there's a bunch of stuff published about that.

Also look at the application guides for low esr high surge electros.

IC has a good blurb

http://www.illinoiscapacitor.com/pdf/Pa ... iction.pdf


So those are the easiest numbers to look at and quantify.

The more difficult aspect to put numbers too is that capacitors have capacitance that is frequency dependent and their esl/esr numbers in relation to their in circuit impedance is frequency and phase dependent.

This means that current draw and circuit load can be highly dependent on input frequency. Electros are the least linear in this regard compared to film.

That's just the tip of the iceberg..but honestly most of the data and plots you can look at don't really matter for guitar amps. But the differences do exist.

The easiest way that I think about it is that there's a schematic of the circuit that exists on paper and there's the real schematic of the circuit that's the actual circuit and unless the drawn schematic is so complex that it's incomprehensible there's going to be a world of difference between the drawn schematic and the actual circuit.

Every component actually needs to be represented by an equivalent circuit...


[img:297:151]http://www.interfacebus.com/Equivalent- ... ematic.jpg[/img]

different categories of components have different numbers for all the external stuff. the more linear and closer to an ideal capacitor the component actually is the stiffer and faster the amp will feel. At least that's been my experience and the usual remarks about amps that utilize film supplies. It's not everyone's cup of tea.

You can get a sense of it by switching between two largely different sized reservoir caps in the power supply. At a certain point the actual value doesn't matter because the caps supply the current. When that happens increasing the capacitance is going to decrease the ESR/ESL and change the values in the equivalent circuit.

An easy thing to do is 47uf and 470uf as the first cap.

Well this is what Andy fuchs is saying on his website about the Ps filtering on the train 45:

Audiophile style circuit techniques bring out the best in any tube circuit. Guitar amps are no different. The Train-45™™ starts with an audiophile style high voltage storage system. Instead of using single large capacitors, the Train-45™ has a group of smaller capacitors in a bank. By using a bank of these smaller capacitors ganged together to produce one large capacitance, you end up with a power supply that is both forceful yet agile. One that holds lots of energy for transients yet can respond quickly to each and every note.

ya that's just the description of the bank. I think it's 4 in series parallel for the reservoir. I'm still trying to find where he explains it. Although, calling it audiophile style is a bit of a stretch...it's not nearly over the top enough to qualify.

this one is my favorite

[img:800:494]http://jimmyauw.com/wp-content/uploads/ ... cb-bgf.jpg[/img]

I think I have a picture of 40 film caps saved around here somewhere.

[talbany]

I used the Solens (banked) in a JTM 45 build and didn't care for them.. Too fast and rather lifeless!!. One ended up in a 18 watt Marshall build that sounded for my tastes wonderful!!

Tony

[Colossal]

I used the Solens (banked) in a JTM 45 build and didn't care for them.. Too fast and rather lifeless!!. One ended up in a 18 watt Marshall build that sounded for my tastes wonderful!!

Tony

Tony,

When you tried Solens in your JTM45, did you substitute Solens in with the same values in every position as the stock amp? In other words, did you use a 32uF Solen for the reservoir cap which calls for 33uF? One for one?

[John_P_WI]

Sometimes putting a small resistance, say 10 ohms or less between the plastic cap (Solens) and ground in the PS can help warm them up and give a more vintage feel. The advantage is the plastic caps will last our lifetime.

[talbany]

I used the Solens (banked) in a JTM 45 build and didn't care for them.. Too fast and rather lifeless!!. One ended up in a 18 watt Marshall build that sounded for my tastes wonderful!!

Tony

Tony,

When you tried Solens in your JTM45, did you substitute Solens in with the same values in every position as the stock amp? In other words, did you use a 32uF Solen for the reservoir cap which calls for 33uF? One for one?

To be honest I don't remember it was a long time ago and was a scratch build.. I do remember Dave Barber recommending them to me since he used them in his amps.. They didn't last long in my 45
John hanging a resistor to ground changes the RC time constant and Dumble does this on some of his MM amps to help soften them a bit..but I am not certain if it changes or alters harmonic content..

Tony

[Colossal]

To be honest I don't remember it was a long time ago and was a scratch build.. I do remember Dave Barber recommending them to me since he used them in his amps.. They didn't last long in my 45
John hanging a resistor to ground changes the RC time constant and Dumble does this on some of his MM amps to help soften them a bit..but I am not certain if it changes or alters harmonic content..

Tony

Thanks Tony. I was just wondering. Polypropylene caps have a much faster response so it might be necessary to lower the amount of capacitance to achieve the same feeling in replicating a classic power supply. John, interesting idea!

[John_P_WI]

Dave and Tony, IF using large Solens poly caps in the power supply and IF placing a resistors between the cap and ground, the resistor can be shunted with a switch negating the resistance. This give a modern / vintage switch feel.

If multiple caps are used in parallel and a larger "switched" resistor 100 to 1000 ohms is used between one of the caps and ground, the cap can effectively be removed from the circuit giving a loose / tight feel too.


John

[diagrammatiks]

I usually go for about half.

For the preamp stages, the SLO only uses 10uf with a 15k droppin resistor.

From an AC filtering stand point that gives you -40 db of ripple attenuation with a -3db frequency of 1 hertz. Decreasing the cap to a 4.7uf there only lowers the ripple attenuation to -34.5 and raises the corner frequency to 2.2hertz, still perfectly acceptable for filtering.