In one of my amps, I have a pi filter followed by a voltage regulator. I have been using some largish resistors to drop lots of volts in the PS as the power transformer is a bit large. I ran some simulations with Duncan Amps designer, and found that using the parts I have, but switching to a choke input PS, I can essentially lose all of the large resistors and replace them with small ones (from 500R to 20R). Aside from taking out the large resistors, it will also lower the voltage dropped across the regulator which seems like a good thing.
My experience so far with this amp is that different rectifiers sound hugely different, and in general, the more volts they drop (and thus the less the regulator drops), the better the amp sounds. I would guess this has less to do with the tubes, and more to do with very hot mosfet regulator.
So, are there drawbacks I should know about here? The choke on the input will be the only choke in the PS, in case that is important.
My experience so far with this amp is that different rectifiers sound hugely different, and in general, the more volts they drop (and thus the less the regulator drops), the better the amp sounds. I would guess this has less to do with the tubes, and more to do with very hot mosfet regulator.
So, are there drawbacks I should know about here? The choke on the input will be the only choke in the PS, in case that is important.
I'm curious as to the topology of the amp. Regardless of that though, I can't see how choke input could ever be a bad choice if you have voltage to spare. I'm interested to see what others have to say on this excellent topic. Thanks, Jay
from what I have read, the only drawback to choke-input are:
1. the lower voltage (according to an article from audioexpress turorials)
2. chokes are more expensive than resistors (my reasoning)
3. size if space is limited (my reasoning)
1. the lower voltage (according to an article from audioexpress turorials)
2. chokes are more expensive than resistors (my reasoning)
3. size if space is limited (my reasoning)
With solid state rectifiers, you would have to think about protecting them from inductive spikes - no need with a tube. You WILL lose some filtering with one less cap, of course, but if you're regulating afterwards, no big deal.
Be sure to size the chokes current capacity such that it can handle the AC ripple current present in addition to its DC load.
VinRMS/1386/L= ma AC ripple 60hz mains.
VinRMS/1386/L= ma AC ripple 60hz mains.
Two issues you might encounter. If the amp takes longer to come up than the power supply, the choke will drop out of the power supply and the voltage will rise as if it wasn't there. Of course this will drop back down once the amp starts drawing current. An appropriately sized bleeder resistor will help to minimise this effect.
Secondly you could experience ringing between the power transformer and the choke. This will probably manifest as overheating of the power transformer. This is a problem I encountered when I used a choke input filter to bring a 900-0-900V transformer down to 700VDC.
Shoog
Secondly you could experience ringing between the power transformer and the choke. This will probably manifest as overheating of the power transformer. This is a problem I encountered when I used a choke input filter to bring a 900-0-900V transformer down to 700VDC.
Shoog
You can check for likely ringing in PSUD II with a stepped load. Reducing ringing involves using more capacitance and/or resistance in the LC filter. Im not sure about ringing in the PT though, this is for ringing in output voltage due to transients.
It's common to use a bleeder resistor to ground after the choke, if the load is variable, to ensure that the choke does its job even with light load. Also, with variable load such as that imposed by class AB or B amplifiers, the choke may need to be a swinging choke (a choke with a smaller than usual air-gap in its core), so that its inductance falls with increasing current to improve regulation.
Swinging chokes tend to be a custom item these days, youl still have better regulation than a CLC filter regardless.
Swinging chokes also dont need as large a bleeder.
Swinging chokes also dont need as large a bleeder.
Thanks for all the comments. I switched the PS around last night and everything seems to work pretty well. Sound wise, I am not sure I noticed much change, though switching from a Sovtek 5ar4 to a Mullard one did seem to smooth things out a bit. But, the whole amp runs cool, other than the tubes of course, but the transformers and top plate are all staying cool.
One potential problem is that ripple before the regulator went from 100uV to about 100mV. (Previously, this was a RCLCRCRC filter before the reg, now it is a LCRCRCRC.) This is a headphone amp, and this was enough to create a very slight audible hum. So, I may need to bump up the filter caps a bit.
-d
One potential problem is that ripple before the regulator went from 100uV to about 100mV. (Previously, this was a RCLCRCRC filter before the reg, now it is a LCRCRCRC.) This is a headphone amp, and this was enough to create a very slight audible hum. So, I may need to bump up the filter caps a bit.
-d
Try placing an 0.47uf cap before the choke. This shouldn't increase the voltage but might help with the ripple.
Glad to hear it worked out.
Shoog
Glad to hear it worked out.
Shoog
Shoog said:
It looks like it will increase the voltage by 10V or so, though this is no bad thing. It allows me to use some other NOS rectifiers I have around. But, the ripple only drops a little - from 78mV to 74mV at high currents and by a bit more (~20mV) at lower currents -- current depends on the tubes I use. But, because of the choke input, the voltage stays low enough that I can use lower voltage caps which should allow a larger cap in the first position (after the choke) with the same footprint, which should help.
- Status
- Not open for further replies.