Sorry, one last PSU question. Overspec caps?
Goober
I'm building a valve project, and the psu calls for caps in the normal
Bridge rectifier - cap1 40uf (filters to earth) - 10k resistor - cap2 10uf
filters to earth)
Will it cause any problems if I substitute the 10uf for another 40uf?
I can't seem to find any 10uf capacitors with over 200v rating, but 40ufs
are easy to find!
The project is an altec 436c compressor.
I can post a schematic, and it's pretty easy to find one on the web as well.
Jim Doyle
: Is it a good idea to uprate smoothing cap values in psu filters?
Uprating the rated working volts is a good idea -- however uprating
the capacitance has implications for the rest of the components.
Particular for the effect of ripple current.
Q = CV take the time derivative i = dQ/dt = C dV/dt
When the rectifier begins conducting, the above equation determines
the current drawn through the rectifier (and xform secondary) as
the cap begins to charge.
doubling the capacitance thus doubles the ripple current...
Make sure that everything else can tolerate the even larger current spikes,
i.e. valve rectifiers, chokes, silicon rectifiers.
-- Jim
Ian Iveson
As far as I know, there could be a danger of low-frequency oscillation, if
you happen to hit on the wrong value. For Hi Fi I would just use something
even bigger if this happened.
But with a compressor, you may find it has a totally different sound. It
may depend on a voltage drop developed across the 10k resistor for some of
its compression. The change of cap would move this effect miles down the
frequency range, depending on the circuit.
So try a 40 and see how it goes. The cap itself will be well protected by
the resistor. Remember that you can make up a value by putting several in
series or parallel. If in series, remember to use balancing resistors
across the caps.
regards, Ian
"Jim Doyle" <do...@world.std.com> wrote in message
news:G96sG...@world.std.com...
Goober
"Ian Iveson" <ianives...@virgin.net> wrote in message
news:dbAl6.17172$MN.4...@news2-win.server.ntlworld.com...
dmot
and thus the amount of compression of sound the tubes can give 'cause they
get more/less juice.
In tube amps psu's the values of capacitors aren't so low cause its
cheaper....
Goober <fu...@wobbly.swinternet.co.uk> schreef in berichtnieuws
zThl6.12072$MN.3...@news2-win.server.ntlworld.com...
Tim
> rectifiers.
500uF!!!? Where in heck'd you get those?
Hmm, what was the V(oltage) and $(us) ?
Though I myself have seen such large caps, MOST
definetly not for such a purpose.
Hey, apply power to just the heaters and reply back
'bout how long it takes to discharge the caps...
Minute? Two?
Or is that a 200W+ amp, thus needing such high values?
No, of course, it can't be; you said they *were* 40uF.
Anyway, compression is (I guess it is anyway) basically
amplifying the signal logarithmically such that loud
passages become quiter and vice-versa.
I don't think the caps have anything to do with that;
of all things they'll do is reduce hum, maybe general noise.
Tim
Ian Iveson
My ppp EL84 domestic amp uses just a 10ohm resistor followed by a 390uF,
450V cap that I got military surplus for 4 UKP a piece. Hence I use the
same cap for the driver and input stages after dropping with about 8k.
No hum whatsoever. No motorboating. But I guess there will be some
roughness on sudden large transients in combination with much bass. That
is, when it is forced into compression then it will not do it kindly. A
choke and smaller cap would be much smoother under such circumstances.
Now a compressor is obviously built to give sweet and soft compression all
the time, rather than as a response to crisis like it is in a domestic hi-fi
system. It deliberately and progressively clips to achieve this, and the
progression of the clipping may well be partly controlled by the caps in
question.
Different horses for different courses.
regards, Ian
"Gee Yip" <ma...@rocketmail.com> wrote in message
news:3A976ADE...@rocketmail.com...
> Hi Tim,
>
> IYes it is two 500uF 450V in a pie configuration with a 10H. It's the
size of a
> coke can. There is a relay that disconnects the B+ when the power is
turned off
> and the cap discharges slowly. When it powers on the relay applies B+
after 30
> seconds or so. B+ of 360V has a 40mV ripple. Kind of crazy right.
>
> I am not sure if it sounds better or not, but after the upgrade it sounds
> different.
>
> It's a EL84 PP. I thought it wouldn't hurt to beef up the power supply.
So
> hence the big cap, plus I have room on my chassis. It cost me around
$20US for
> each. I bought it on ebay. I still have 2 more. I bought 4.
>
> So the beefy B+ won't affect compression, right?
>
> Thanks,
>
> Gee
>
> Tim wrote
dmot
Gee Yip <ma...@rocketmail.com> schreef in berichtnieuws
3A96F9FA...@rocketmail.com...
> Hi,
>
> Can you elaborate on this thingy call compression, please. What does
increasing
> the
> cap have to do with sound of the tube.
I don't recall exactly, but when you use lower values caps, the caps get
drained faster.
That makes the tubes get less power, so power decreases. That way, you'll
get less clipping
when the tube is pushed to its max.
Tube clipping is the nicest cause there's soft cut-off point, transistor
clipping sounds harsh, because
there's no round cut-off point.
When using larger caps, you'll get more of a "transistor clipping"
please don't nuke me on this, I'm not a tube guru..
greetings dmot
Ned Carlson
<ianives...@virgin.net> wrote:
>But with a compressor, you may find it has a totally different sound. It
>may depend on a voltage drop developed across the 10k resistor for some of
>its compression. The change of cap would move this effect miles down the
>frequency range, depending on the circuit.
Here's the diagram for the unit he's referring to:
http://www.triodeel.com/al436c.gif
and the PS voltage going to the tube that does
the compression (6BC8) goes UP about 80 volts
when the compression is cranked, not the other way around!
Actually, I'd imagine that replacing the 10K resistor with
a small choke and using a bigger capacitor might help
prevent this unit from pumping under heavy signals.
Ned Carlson Triode Electronics "where da tubes are!"
2225 W Roscoe Chicago, IL, 60618 USA
ph 773-871-7459 fax 773-871-7938
12:30 to 8 PM CT, (1830-0200 UTC) 12:30-5 Sat, Closed Wed & Sun
<A HREF="http://www.triodeel.com">http://www.triodeel.com</A>
Tube and Tube Amp info on the net...<A HREF="http://www.triodeel.com/tlinks.htm"> The Big Tube Links Page!</A>
Ian Iveson
I don't know what the pots are for, in terms of their effect on the sound.
It appears that compression is obtained by cutting off rather than
saturation, which is not what I had expected. I see that there is positive
feedback via the full-wave rectifier, and that the DC voltage at the first
grids is determined by P3 and P2. The incoming signal is controlled via P1
and P2, and the feedback is controlled by all three pots? Presumably the
feedback accentuates the curve as it approaches clipping, and stops it from
bottoming out. I can't work out why P3 is connected to the cathodes of the
output valve, which is otherwise conventionally employed.
Can you explain what you mean by pumping? I assume this is the integrating
effect of the cap and 10k combination, so when compression happens the
effective dc component on the signal is added to the bias slowly as the cap
finds the new DC level. So a bass flourish will turn down a quite treble
part, which will then take time to recover when the bass stops?
But anyway, stumbling in the dark, it seems that replacing the 10k resistor
with a choke will stop the ratchet by allowing the cap to charge and
discharge without the delay of the resistor. But the DC operating point of
the valve could change drastically, surely? How would Goober decide on the
spec of the choke to give the same voltage? And will the change effect the
attack of the signal?
What is clear, though, is that the primary function of the cap is just
smoothing, and apart from the pumping, it has no effect on the compression,
because there is very little current flowing then.
cheers, Ian
"Ned Carlson" <postm...@triodeel.com> wrote in message
news:3a9854fd....@news-west.newscene.com...
Goober
The compressor is not so affected by the power supply, as it relies the
special variable mu chareteristic of the 6cb8, which changes its gain
dramaticly more than a normal valve when the operating point is moved by the
changing of the grid volts. It does not clip, that would cause distortion,
which is not a good idea in this case! It should be totally clean.
P1 is the input level.
P2 is the compressor threshold.
P3 is the release time. (It discharges c4)
M1 shows the amount of compression.
The rectified output signal is used to change the operating point of v1.
I will build a proper regulated PSU next time, but my lack of design
experience made me just go for the easy solution.
One problem, I've built the damm thing, with a 40uf instead of the 10uf in
the psu (and getting rid of the voltage doubler and replacing it with 200vac
from the transformer and a bridge rectifier with another 40uf after it).
Plugged it in, turned it on. Nothing.
The HT works fine,turns out that the heater secondary on the transformer is
bad.
Damm! Have to wait a few weeks for a replacement.
"dmot" <dm...@gmx.net> wrote in message news:97866c$431$1...@news.hccnet.nl...
Ian Iveson
regards, Ian
"Goober" <fu...@wobbly.swinternet.co.uk> wrote in message
news:_Ldm6.26483$5n4.5...@news6-win.server.ntlworld.com...
Goober
incompetent soldering, not the transformer :)
Great compression, nice warm sound.
The only problem is, it does 'moterboat' with tickthumps that start when the
input volume is about half way, and get faster as it is advanced. It's not
too much of a problem, as I can drive it harder from the mixing desk to
compensate, but would be nice to cure.
I'll try replacing that psu cap, but should I go bigger or smaller?
Ian Iveson
least I don't get the feedback thing. It appears to be -ve and +ve
alternately. The rectifier valve doesn't just provide the bias. How well
does your release time pot work?
Warmth is harmonic distortion in the right proportions. I have been
pondering how such a circuit performs wrt IMD?
If I could simulate it I could do a frequency analysis I suppose on the PS.
But loads of people here will know. Perhaps you might start a new thread if
we are forgotten down here? But are you sure it's motorboating? If so then
I think the idea is to drop the resonant frequency, which would mean a
different resistor/cap combination. Either a bigger cap or a smaller
resistor I suppose? Just stirring up thoughts here.
cheers, Ian
"Goober" <fu...@wobbly.swinternet.co.uk> wrote in message
news:EtVm6.3025$OL.1...@news6-win.server.ntlworld.com...
Goober
> I can't help you with that because I still don't know how it works! Or at
> least I don't get the feedback thing. It appears to be -ve and +ve
> alternately. The rectifier valve doesn't just provide the bias. How well
> does your release time pot work?
The output signal is rectified, and then smoothed by that release
resistor/cap combination, providing a slow control voltage for v1, which
acts as a VCA. (Someone with more experience please correct me if i'm
wrong!)
>
> Warmth is harmonic distortion in the right proportions. I have been
> pondering how such a circuit performs wrt IMD?
Well, I fired up the old spectralab, and frankly the IMD sucks! Lots of
harmonics, mostly related, and a little bit of PSU hum breaktrhu.
It's not a nasty kind tho, and seems to interact in the HF with the
threshold and release controls. No idea why, even though its not that big a
circuit, I think there is quite a bit off odd interaction between the
controls.
Doing a few listening tests, it seems to work well of vox and steel acoustic
guitar.
The frequency response is pretty good, I can't see much difference between
white noise loopback and the compressor, just a slight roll-off above 18k,
probably the input transformer.
>
> If I could simulate it I could do a frequency analysis I suppose on the
PS.
> But loads of people here will know. Perhaps you might start a new thread
if
> we are forgotten down here? But are you sure it's motorboating? If so
then
> I think the idea is to drop the resonant frequency, which would mean a
> different resistor/cap combination. Either a bigger cap or a smaller
> resistor I suppose? Just stirring up thoughts here.
I'll try it. It's fairly easy to swap, and cos I thought I might have
problems here, I made it accessable!
FT
surprising that the HD & IMD would be quite high. One more reason
to use these as little as possible, if you're interested in high
fidelity. If you're shaping the sound of your guitar or bass amp,
then go for what you like.
.
Pumping is a kind of low frequency oscillation of gain that happens
through the feedback of the compression process itself, rather than
via an inadequate power supply. If you give the compressor too much
signal it turns the gain sharply down, and then the resulting
sharply reduced signal causes the compressor to turn the gain
sharply back up, and off we go. Careful attention to the time
constants in the feedback loop will help minimize this, but any
analog compressor will pump if pushed hard enough. Actually, I
suspect any analog compressor will always be pumping to one degree
or another if it has a changing signal going through it, whether
it's noticeable or not.
IMHO, this kind of signal processing is one of the areas where
modern digital processing will usually win out over tubes. If
you're going to sacrifice the integrity of the original signal by
running it through a compressor, you've just left most of the
advantages of tubes behind, unless you're out to create a certain
distorted sound in the first place.
Fred
"Goober" <fu...@wobbly.swinternet.co.uk> wrote in message
news:9sYm6.4025$OL.1...@news6-win.server.ntlworld.com...
Ian Iveson
point that a rectified signal will charge up the cap. But there is still
some remains of that signal on the grids of the first valve, and is negative
feedback as the signal goes negative, and vice-versa. Perhaps not
significant? I can imagine this effect with a sine wave, and simulate it.
But with a more complex signal it gets hard to grasp what is going on.
Anyway, perhaps Fred has identified the problem oscillation at high
compression levels? I guess any changes to any part of this circuit will
need careful compensation elsewhere, because of the interactive controls,
and the fact that you have two different time-related processes going on at
once in the same circuit.
Good to learn something, thanks.
Ian
"Goober" <fu...@wobbly.swinternet.co.uk> wrote in message
news:9sYm6.4025$OL.1...@news6-win.server.ntlworld.com...
Ian Iveson
valves, I now see how it should work.
P3 controls the current to C4, and P2 controls the voltage. So P3 changes
the speed of charging C4, via pulse-width-modulation, kind of, with P2
controlling the peak voltage. This is a compromise, isn't it? On the one
hand you want to filter out the rectified and chopped audio signal, and on
the other you want to control the time constant to give an acceptable
release time. No doubt the apparatus has sweet spots where all this comes
out in the wash.
I suggest the input valve is cutting off when you get that noise? Have you
checked all the voltages? And I agree that all the time constants are
interactive and quite critical.
cheers, Ian
Goober
> the speed of charging C4, via pulse-width-modulation, kind of, with P2
> controlling the peak voltage. This is a compromise, isn't it? On the one
> hand you want to filter out the rectified and chopped audio signal, and on
> the other you want to control the time constant to give an acceptable
> release time. No doubt the apparatus has sweet spots where all this comes
> out in the wash.
the pots, but it could be cos I am seeing it in audio engineer terms rather
than electrical!
P3 sets the threshhold, it adjusts the level that if crossed, compression
occurs. P2 sets the release time. The release time varies from about 0.2
seconds to about 2 seconds. I think r12 sets the attack time of the
compressor. I might put this on the front panel later. R9 is 270k by the
way, but I got this from the altec 436b diagram, as It is not marked on the
436c schematic, perhaps it needs to be selected seperately for each
individual compressor? This could be perhaps where I am going wrong.
It does not matter if there is audio in the control signal that goes to v1,
as because it would be applied equaly to both sides of v1, it will be
cancelled out by t2. The thump of the control voltage is cancelled out in
the same way. The only thing you should hear is the audio, as it is out of
phase in both halves of v1. In practice, cos the gains of each side of v1
and v2 are not exactly matched, you do get a little bit of it coming thru.
> I suggest the input valve is cutting off when you get that noise? Have
you
> checked all the voltages? And I agree that all the time constants are
> interactive and quite critical.
How do you mean cutting off? Do you mean it's running out of headroom?
The slow rhythmic clicking/thumping happens when there is no signal going
through the device at all, and starts and beginnins to get faster as the
input gain control (p1) is advanced from about 1/2 way. When you put a
signal through, it stops happening.
Also, the meter (m1) goes sharply to the left in time with the clicking.
When the gain control is at maximun it becomes almost a low bass note around
50 hz. The threshold control makes the clicks less thuddy, and the release
time makes larger gaps between the clicks. With all controlls on max, the
clicking stops, or is inaudible, but the meter reads a steady 1/2 way
compared to the reading when there is no clicking ( with input gain low).
I have not checked the voltages yet, as I am still getting over the shock of
the thing working at all :)
I agree with fred about the oscillation being due to the feedback chain, it
seems to make perfect sense, but I'm not really sure what to do about it,
maybe increasing r12?
There seems to be more than usual distortion on low frequency signals, you
expect this at some settings with a compressor, but the time constants in
this one are way to long to cause it, so it must be something else!
Andy Cowley
Dear Gee,
Valve amps compress the signal because the B+ falls when the valves
draw large currents on peaks. That is the impedance of the power supply
is high. Bigger caps lower the (transient) impedance so you get less B+
collapse and a different sound. Not quite the same thing that a compressor
does but a much sought after sound for instrument amps.
Andy Cowley
p.s. you could A/B this by switching between big caps and little ones,
bypass the switch with a few hundred K to keep all the caps charged.