I plan on building a power supply unit housed in a separate enclosure by itself to power a mono F5V3.(there will be two amp units and two power supply units) I will be building the PSU based on the schematic for the F5V3 article and will be using a transformer with 34 V secondaries. I need help with the following:
1) Since I will have to use an umbilical between the two units what kind of connectors should I use to connect the power unit to the amp unit?
2)I have read that the power capacitor bank should be as close as possible to the amplifier board. Should I mount the capacitors in the amplifier unit(I do have space for them) and have the PSU contain only the transformer and rectifiers and feed the umbilical from the rectifier outputs to the caps in the amp unit, or, should I build the PSU complete with the caps and feed the output from the cap board to the amplifier section with the umbilical? Whats the better approach?
All advice will be appreciated. Thanks.
1) Since I will have to use an umbilical between the two units what kind of connectors should I use to connect the power unit to the amp unit?
2)I have read that the power capacitor bank should be as close as possible to the amplifier board. Should I mount the capacitors in the amplifier unit(I do have space for them) and have the PSU contain only the transformer and rectifiers and feed the umbilical from the rectifier outputs to the caps in the amp unit, or, should I build the PSU complete with the caps and feed the output from the cap board to the amplifier section with the umbilical? Whats the better approach?
All advice will be appreciated. Thanks.
I disagree.
The smoothing capacitors should be close to and electrically compact with the transformer secondaries and the rectifiers.
The electrical compactness of the PSU is achieved by using short connections/wires and using low loop area for all flow/return connections.
This electrical compactness is necessary to minimise radiated interference.
There was a Thread discussing this separated Chassis issue a few years back, maybe around 2008/9
Safety proved to be one of the biggest issues in that discussion.
There is a current Thread discussing this and Safety seems to be taking a back seat behind hum reduction.
Thanks Andrew. Your explanation is a great help to a newbie like me. I will then make the PSU with the transformer, rectifiers and cap board all in one chassis with as short wires as possible connecting these components together. Since the umbilical connecting the power chassis to the amp chassis will be about a foot long will the distance pose any issues?
The schematic of the amp board shows decoupling caps C1 and C2 are 10uf. Since I am going the separate chassis route do you feel I should change the value of these caps? Any other suggestions?
I searched but could not find the separated chassis Thread you are referring to.
Thanks.
Thanks Andrew. Based on this I will add a Safety ground between the PSU chassis and the amp chassis.
What is the .0033 capacitor in Nelson's F5 Power Supply schematic for?
In the present Pass lineup the XS series has 2 boxes per channel(amp unit and PS box) but I wonder why the PS box has what appears to be the same amount of heatsinking as the amp unit?
nashbap, it’s a Very interesting question …. Particularly in the case of the F5Tv3 and its CRC PSU.
Some rational lateral thinking should provide a best-of-all-worlds, unique solution.
I agree with the approach of having large capacitance adjacent to the output devices power bus. It will certainly lead to better transient current delivery, at high output power. Also, since the F5Tv3 does not run to full power in Class A, transient power will certainly be called for, by the amplifier. A close proximity (to the output devices) capacitor bank will be beneficial under such conditions.
The most important negative that comes to my mind of locating capacitors in the Power Amp housing is the high ambient temperature of the power amp chassis. A capacitor’s life is down 50% (in geometrical progression) for every additional 10 deg C in the capacitor’s CORE temperature beyond 70 deg C. That said, a Professional Grade capacitor ( typically ones with screw terminals) have an adequate MTBF (Mean Time Before Failure) to accommodate a fair amount of derating.
I would advice against lodging only the Transformer and rectifier in the PSU Housing. Running a high current 120Hz or 100 Hz rectified waveform will certainly cause a large amount of radiated interference, completely negating the physical isolation of a separate PSU housing.
However, the PSU for the F5Tv3 actually has a CRC ( Capacitor – Resistor – Capacitor ) power supply.
I believe it would be an IDEAL solution to house the Transformer + Recifier + 1st capacitor bank in the PSU housing and the parallel Bank of resistors along with the second capacitor bank in the Power Amp chassis. You can ( should ) locate the capacitors as close as possible to the Power Rail of the output devices.
In the 2 physically distanced capacitor banks, the Resistance ( and inductance ) of the Unbilical will for part of the “R” in the CRC network. Any small added inductance of the Unbilical will not have a negative effect on the CRC performance…. And could arguably even improve it !
Regarding your query on the appropriate connector to be used, its probably most practical; to use any High current connector that is easily available in small quantities. Several MIL connectors would be ideal ( as used in the Rowland 8T) but may not be available in very small quantities. Do explore high current Automotive connectors
However, since this is a DIY project, its worth keeping in mind that NO Connector is as good as a well Soldered connection. As a DIY project, you have the luxury option of eliminating the connector altogether, with hard soldering at both ends.
I am not sure if the 10 mfd cap is an electrolyte cap. I don’t think you need to increase the value of the 10 mfd capacitor though I don’t see any negative in even doubling it. ( Electrolytic caps usually have a -20% to +50 capacitance value tolerance any way ) Papa Nelson Pass is VERY thorough & would have mentioned if there was a need
You could bypass the 10 mfd electrolyte ( if it is an electrolytic) with a 0.1 MFD polypropylene and another 0.001 mfd film ( Stryo or polyprop ) cap to provide High Frequency decoupling. A generic electrolytic cap typically turns into an inductance at frequencies as low as 100 KHz !
Some rational lateral thinking should provide a best-of-all-worlds, unique solution.
I agree with the approach of having large capacitance adjacent to the output devices power bus. It will certainly lead to better transient current delivery, at high output power. Also, since the F5Tv3 does not run to full power in Class A, transient power will certainly be called for, by the amplifier. A close proximity (to the output devices) capacitor bank will be beneficial under such conditions.
The most important negative that comes to my mind of locating capacitors in the Power Amp housing is the high ambient temperature of the power amp chassis. A capacitor’s life is down 50% (in geometrical progression) for every additional 10 deg C in the capacitor’s CORE temperature beyond 70 deg C. That said, a Professional Grade capacitor ( typically ones with screw terminals) have an adequate MTBF (Mean Time Before Failure) to accommodate a fair amount of derating.
I would advice against lodging only the Transformer and rectifier in the PSU Housing. Running a high current 120Hz or 100 Hz rectified waveform will certainly cause a large amount of radiated interference, completely negating the physical isolation of a separate PSU housing.
However, the PSU for the F5Tv3 actually has a CRC ( Capacitor – Resistor – Capacitor ) power supply.
I believe it would be an IDEAL solution to house the Transformer + Recifier + 1st capacitor bank in the PSU housing and the parallel Bank of resistors along with the second capacitor bank in the Power Amp chassis. You can ( should ) locate the capacitors as close as possible to the Power Rail of the output devices.
In the 2 physically distanced capacitor banks, the Resistance ( and inductance ) of the Unbilical will for part of the “R” in the CRC network. Any small added inductance of the Unbilical will not have a negative effect on the CRC performance…. And could arguably even improve it !
Regarding your query on the appropriate connector to be used, its probably most practical; to use any High current connector that is easily available in small quantities. Several MIL connectors would be ideal ( as used in the Rowland 8T) but may not be available in very small quantities. Do explore high current Automotive connectors
However, since this is a DIY project, its worth keeping in mind that NO Connector is as good as a well Soldered connection. As a DIY project, you have the luxury option of eliminating the connector altogether, with hard soldering at both ends.
I am not sure if the 10 mfd cap is an electrolyte cap. I don’t think you need to increase the value of the 10 mfd capacitor though I don’t see any negative in even doubling it. ( Electrolytic caps usually have a -20% to +50 capacitance value tolerance any way ) Papa Nelson Pass is VERY thorough & would have mentioned if there was a need
You could bypass the 10 mfd electrolyte ( if it is an electrolytic) with a 0.1 MFD polypropylene and another 0.001 mfd film ( Stryo or polyprop ) cap to provide High Frequency decoupling. A generic electrolytic cap typically turns into an inductance at frequencies as low as 100 KHz !
The most important negative that comes to my mind of locating capacitors in the Power Amp housing is the high ambient temperature of the power amp chassis. A capacitor’s life is down 50% (in geometrical progression) for every additional 10 deg C in the capacitor’s CORE temperature beyond 70 deg C. That said, a Professional Grade capacitor ( typically ones with screw terminals) have an adequate MTBF (Mean Time Before Failure) to accommodate a fair amount of derating.
I would advice against lodging only the Transformer and rectifier in the PSU Housing. Running a high current 120Hz or 100 Hz rectified waveform will certainly cause a large amount of radiated interference, completely negating the physical isolation of a separate PSU housing.
However, the PSU for the F5Tv3 actually has a CRC ( Capacitor – Resistor – Capacitor ) power supply.
I believe it would be an IDEAL solution to house the Transformer + Recifier + 1st capacitor bank in the PSU housing and the parallel Bank of resistors along with the second capacitor bank in the Power Amp chassis. You can ( should ) locate the capacitors as close as possible to the Power Rail of the output devices.
In the 2 physically distanced capacitor banks, the Resistance ( and inductance ) of the Unbilical will for part of the “R” in the CRC network. Any small added inductance of the Unbilical will not have a negative effect on the CRC performance…. And could arguably even improve it !
I am not sure if the 10 mfd cap is an electrolyte cap.
Thanks for your detailed and thoughtful reply.
I had not considered the effect of radiated interference when I had originally considered the idea of a PSU with only the transformer and rectifiers. As both you and Andrew have said it makes sense to include the cap bank in the PSU. Your idea of separating the second cap bank of the CRC seems like it could be workable. As you mentioned I am really concerned about the impact of heat on the caps even though I did not mention it in my original posting. In the amp unit that I will be constructing there are essentially two large heatsinks separated by a distance of about 5 1/2 inches so the temperature between them can easily be around 50C under operation, maybe more. Even 105C rated caps might be short lived.
Perhaps I could build the PSU with the CRC bank included and maybe at a later date add a couple of caps in the amp unit making it a CRC R(wire) C. Does this make sense?
The 10uf is an electrolytic.
I would advice against lodging only the Transformer and rectifier in the PSU Housing. Running a high current 120Hz or 100 Hz rectified waveform will certainly cause a large amount of radiated interference, completely negating the physical isolation of a separate PSU housing.
However, the PSU for the F5Tv3 actually has a CRC ( Capacitor – Resistor – Capacitor ) power supply.
I believe it would be an IDEAL solution to house the Transformer + Recifier + 1st capacitor bank in the PSU housing and the parallel Bank of resistors along with the second capacitor bank in the Power Amp chassis. You can ( should ) locate the capacitors as close as possible to the Power Rail of the output devices.
In the 2 physically distanced capacitor banks, the Resistance ( and inductance ) of the Unbilical will for part of the “R” in the CRC network. Any small added inductance of the Unbilical will not have a negative effect on the CRC performance…. And could arguably even improve it !
I am not sure if the 10 mfd cap is an electrolyte cap.
Thanks for your detailed and thoughtful reply.
I had not considered the effect of radiated interference when I had originally considered the idea of a PSU with only the transformer and rectifiers. As both you and Andrew have said it makes sense to include the cap bank in the PSU. Your idea of separating the second cap bank of the CRC seems like it could be workable. As you mentioned I am really concerned about the impact of heat on the caps even though I did not mention it in my original posting. In the amp unit that I will be constructing there are essentially two large heatsinks separated by a distance of about 5 1/2 inches so the temperature between them can easily be around 50C under operation, maybe more. Even 105C rated caps might be short lived.
Perhaps I could build the PSU with the CRC bank included and maybe at a later date add a couple of caps in the amp unit making it a CRC R(wire) C. Does this make sense?
The 10uf is an electrolytic.
you must include MF and HF decoupling at the main current consumers in the power amp stages. This could mean a dozen or so decoupling capacitors for one channel of F5t.
Once this is done properly there may turn out to be not much need for LF decoupling next to the current consumers and that this could be located in the coolest part of the Power Amplifier.
This is effectively the final stage smoothing of an rC......RC supply.
Once this is done properly there may turn out to be not much need for LF decoupling next to the current consumers and that this could be located in the coolest part of the Power Amplifier.
This is effectively the final stage smoothing of an rC......RC supply.
I have been trying to read up as much as I can about decoupling but there is no lack of conflicting and confusing information. Given that I will be using a separated supply with 45 V rails could you advice some specific cap values at the amp board to implement the above. I think what you mean by providing MF and HF decoupling is paralleling caps of different values at locations C1, C2 and C5, C6 in the schematic. Nelson is suggesting 10uf at these points but that is assuming a small loop area and not the substantially longer one that results with a separated supply. Thanks.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.