Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

any limit on reservoir capacitance?
any limit on reservoir capacitance?
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools Search this Thread
Old 24th December 2005, 11:16 AM   #11
forr is offline forr  France
diyAudio Member
forr's Avatar
Join Date: Dec 2004
Location: Next door
Hi !

"I think you need to read the earlier posts of ALW and Jan Didden about this. Basically, toriodal tranformers are wideband devices - they can resonate in the MHZ region (I think ALW measured this).
Increasing the capacitance can shorten the conduction angle, and consequently send out much higher frequency trash which is much harder for the amplifier to reject."

I am a bit in difficulties here. I often read that short conduction makes a lot of parasitic trash and long duraction less, however I do not understand the mechanism. I have been looking for an explanation since quite a long time and read ALW and Didden posts but it is still absolutely unclear to me .

Can you help ?

~~~~ Forr

  Reply With Quote
Old 24th December 2005, 11:16 PM   #12
Johan Potgieter is offline Johan Potgieter  South Africa
diyAudio Member
Join Date: May 2005
Location: Pretoria, South Africa
Just about everything has been said, and I agree that I do not see the advantage. But I miss reference (unless I am asleep - don't answer) to low frequency power supply impedance. High value capacitors have more advantage there than to reduce ripple. It depends on what load the amp is required to drive at which (very low) frequency - in other words, a question of regulation. Was this part of the quest?

Also one must not overlook that any power transformer has a finite impedance (wire resistance plus inductance) which will eventually limit the charging current. Maximum current still has to be checked to be within specs, naturally.

Forr, you specifically put your question to Kevinbd, who I am confident can reply quite ably, but while I am pounding buttons (and not having read Jan's and ALW's comments):

Depending on the current drawn at any moment, the larger the capacitor, the higher the energy needed to "restore" what charge has been withdrawn while the diode was not conducting. Also, because the capacity is large, discharge would not have been such that the charge period (also expressed in terms of that part of the mains frequency angle during which charge is restored) is very long. The resultant charging "spike" can thus become comparatively short and sharp, and the spectrum analysis of such a spike will reveal more high frequency components the shorter the duration and higher the amplitude. These could certainly go into the r.f. region and typically manifest as interference induced in surrounding circuitry.
This as a brief preliminary explanation.

Regards all.
  Reply With Quote
Old 26th December 2005, 12:11 PM   #13
forr is offline forr  France
diyAudio Member
forr's Avatar
Join Date: Dec 2004
Location: Next door
Hi Johan Potgieter,

Thanks for the explanation. Took a couple of minutes to fully understand. Very simple in fact. Here how I interpret it.
If a constant current is drawn from a reservoir, diodes supply more current in a shorter time than when the reservoir is smaller, then its impedance higher, and conduction greater. Hence the trash with big capacitors.

A second question I aksed myself since 18 years !
In an interview published in Hifi News, july 1987, Stan Curtis said :
"[...] I set out with Rotel to produce a musical amplifier with all the virtues of good focus, good depth, reasonable transparency, but with a very tight, precise, indeed dry, bottom end. I used a trick, which is a valid trick, of tuning the power suplly, which is a matter of adjusting the parameters of the transformers, rectifiers and capacitors, such as it was able to give the maximum transfer of energy form the power supply to the output stage in the bass guitar frequency region, and that meant that when bass guitar was playing, the sound appeared to be very quick, because there was very quick transfer of energy through the amplifier and into the speakers. Below that region, the power suplly got into trouble, and the sound became flabby; above that, the power supply became less dependant on the characteristics of the reservoir capacitors."

Later, Curtis has often refered to this idea. In Electronics World, februrary 2005, he wrote :
"I designed a hugely successful series of amplifiers for Rotel which, despite having a reasonably flat frequency response, did emphasise the bass guitar and bass drum sound so that a 'foot stapping' sound was produced. And tap their feet was the public did whilst buying the products in hundreds of thousands".

Any idea of how to repeat the trick with more scientific bases ?

~~~~~~ Forr

  Reply With Quote
Old 26th December 2005, 12:20 PM   #14
lt cdr data is offline lt cdr data  Wales
diyAudio Member
Join Date: Dec 2003
Location: earth
And I read that Julian Vereker, deceased of Naim, said that he 'designed' (or used the RCA circuit?) on the NAB 300, one of the prototypes

'with bigger time constants, to put a hump in the bass response.'

In other words, a marketing/ way of making himself look more technical when he simply could have said bigger capacitors?

Interested to know which amps Stan Curtis did for Rotel.

Just out of interest, I have paper copies, sadly can't post them here, of Rotels 970 integrated, and lost it hopefully temporary, rotels 1060 integrated amp.

the topology is identical, iirc, the only areas that were changed were compensation caps., and the michi power amp is very similar, as is the 1090 380watt monster amp.

so there is a very tight lineage from at least the 970, probably 971, 1060 and 1062 integrated amps which use near identical circuits with very minor tweaking indeed, I am quite interested in how a maker like this refines a design.
  Reply With Quote
Old 26th December 2005, 02:36 PM   #15
ilimzn is offline ilimzn  Croatia
diyAudio Member
Join Date: Feb 2005
Location: Zagreb
It surprises me that no-one is mentioning two other mechanisms by which increasing filter caps (all else being equal) may be a problem.

The first is an increase of peak current through the rectifier, due to shorter conduction angles. Ultimately, this is limited by the finite impedance of the transformer and wiring - and this is fortunate. With higher peak currents, diode recovery time increases as more excess carriers have to be removed in order for the diode to stop conducting. Although this is not as big a problem in a mains power supply, any switced mode PSU designer will tell you that unchecked diode reverse recovery wastes a lot of heat on the diodes (and transformer) and generates all sorts of HF hash. In other words, if you want to increase the capacitances, you need to think about faster recovery diodes - this helps the losses, which you will see as diodes heating up to unexpected levels. Secondly, soft recovery helps to reduce HF hash, though you will still get more HF energy due to the conduction angle being shorter (i.e. think of it as a higher frequency fundametal, so the harmonics also extend higher).

The second mechanism is more inocuous, and again has to do with higher peak current. It is no accident that you cannot normally have 100W of rectified and filtered AC out of a 100VA transformer - the reason is that the conduction angle of the diodes is not the full sine wave - which means, your output current is supplied in pulses, and to satisfy the power demand, the pulses have to be the higher the shorter they are. On the other hand, the pulses correspond to magnetic flux in the core - and yes, you can quite easily get into asturation, unless your transformer is overdimansioned. Every time the core saturates, the 'excess' magnetic field is not contained within it any more and you get a means to efficiently impress the 'tops' of the pulses onto pretty much everything, by means of inductive coupling. To prevent this, transformers are normally enclosed in a shorted winding of sorts, that wastes a large part of the stray magnetic field as heat. While the transformer is saturated, it does not behave as a transformer any more, and the primary acts more or less as a resistor, rather than an inductor (actually, it's inductance falls several orders of magnitude and the resistance becomes dominant). And, finally, higher harmonic presence in the core increases core losses, All in all, you get a transformer that potentially generates a lot more heat and transfers less power to the load. Not a nice scenario!

As far as inrush current, there are NTC inrush dampers, a good place to look for them are old PC power supplies, if you can't buy them. ALternatively, you can build an inrush current limiter, usually a resistor and relay is all that is required. But, this may be a relatively easy problem considering the above.

Finally, even though going from 10000uF to 27000uF is excessive, using 12000 or 15000uF may not be. The reason is that older caps tended to deliberately be specified with a large + tolerance, to account fo aging, so it is not uncommon for a large filter cap to have +50% and -10% tolerance. Today the tolerances are much narrower as electrolytic technology has advanced considerably. These days if you buy a 10000uF filter cap, it is likely to be within +-5% as taken off the shelf, so you may go up to 15000uF if the original had wide tolerances. This is again, all other things being equal. What you may want to look into is low ESR/ESL. Unless your originals were really bad, even using low ESR/ESL caps will not limit the capacitance, unless the originals had so high ESR as to be on the order of the resistance of the transformer windings - but that is unlikely. Low ESR will shorten the conduction angle of the rectifiers by a small fraction, but will present a signifficanlty lower impedance at audio frequencies, which is important for the amplifier. If you have a choice in the matter and it does not increase the price unreasonably, it is always worth investing in lower ESR capacitors even of the same capacitance.
  Reply With Quote
Old 26th December 2005, 02:42 PM   #16
poobah is offline poobah  United States
diyAudio Member
poobah's Avatar
Join Date: Nov 2005

It makes me laugh when amp designers have to use the all the silly words only because people expect to hear them... "fast bass"

I think all this means low output impedance and high damping factor for the amplifier...
  Reply With Quote
Old 26th December 2005, 03:25 PM   #17
jackinnj is online now jackinnj  United States
diyAudio Member
jackinnj's Avatar
Join Date: Apr 2002
Location: Llanddewi Brefi, NJ
any limit on reservoir capacitance?
Originally posted by Kevinbd
I think you need to read the earlier posts of ALW and Jan Didden about this. Basically, toriodal tranformers are wideband devices - they can resonate in the MHZ region (I think ALW measured this).
It's not the transformer which oscillates it's the combination of the leakage inductance of the transformer PLUS the junction capacitance of the rectifiers (and all of the other associated reactive components). I have observed the oscillation anywhere from 25kHz to a few hundred kHz. (and the oscillation is problematic for reasons we shouldn't get into without risk of thread-hacjubg) Nonetheless, an RC network on the diode bridge may be necessary to prevent diode-ringing.

Now, getting back to the principal topic, if you increase the capacitance the voltage droop will be lessened so the amplifier may be capable of higher output voltage (notice I stated this in the subjective). With higher voltage of course the output transistors will run hotter -- and this isn't trivial in a commercial design in which the heat sink size has been optimized for thermal impedance vs cost. The power dissipated by the heat sink is a function of the square of the supply voltage. In non-switching amplifiers energy is used by the speakers AND to keep your room warm.

Assuming 4 amps, nominal +/- 40V, t = 0.0075S, going from 0.01F to 0.033F will reduce the voltage droop by about +/- 2 volts on the power supply. The heat sink required has a thermal impedance which is lower by roughly 11%. Your mileage may vary, read the prospectus carefully before you invest, known to cause cancer in laboratory rats, etc., etc.
  Reply With Quote


any limit on reservoir capacitance?Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
PSU Reservoir Capacitors (Single large vs Multiple small) Deofol Solid State 15 13th March 2009 06:03 AM
reservoir capacitance for power output lt cdr data Solid State 0 7th July 2007 02:14 PM
53v limit due to caps ? snkby Class D 9 14th January 2007 04:11 PM
Capacitance Limit? Christian Solid State 36 10th January 2003 02:09 AM
100Kb Image Limit = TOO LOW ajhawkins Everything Else 1 3rd January 2002 10:08 AM

New To Site? Need Help?

All times are GMT. The time now is 07:39 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 16.67%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio