I am going to be building a new set of Class D amplifiers for our 5.1 Active EQd system. This will require four stereo amps, and more if we go from 2 way to 3 way speakers on our fronts. Will probably use the HiFiMeDIY T4 amps.
In the HiFi MeDIY T4 thread we have been discussing advantages of using separate power supplies for each amp and also using way over-sized supplies for improved SQ.
I am now seriously considering having all amps in one case, with room for more if needed in the future. The supplies would be housed in a separate case. This case would also be over-sized. I can then do some experimenting with different size and quantities of supplies without messing up the amp case. It would also keep AC out of the amp case and possibly allow for shorter signal leads within the amp case. Wiring between the two cases would of course have to be more than ample to handle the currents and to minimize IR drop.
Does anyone see any disadvantages to this approach? Any suggestions for implementing it?
Thanks,
Rod
In the HiFi MeDIY T4 thread we have been discussing advantages of using separate power supplies for each amp and also using way over-sized supplies for improved SQ.
I am now seriously considering having all amps in one case, with room for more if needed in the future. The supplies would be housed in a separate case. This case would also be over-sized. I can then do some experimenting with different size and quantities of supplies without messing up the amp case. It would also keep AC out of the amp case and possibly allow for shorter signal leads within the amp case. Wiring between the two cases would of course have to be more than ample to handle the currents and to minimize IR drop.
Does anyone see any disadvantages to this approach? Any suggestions for implementing it?
Thanks,
Rod
An idea I had for a friend of mine who wants to build a 5.1 sys (actually it still is an idea), was to keep vertically T4 boards, in order to have more space inside the case, leaving place enough for PSUs. IMHO it's not so useful keeping away AC and PSU... my implementation uses shielded cables and so on (shield linked to earth), but also the board seems to be immune to EMI: I've tried to make a call with my mobile on the heatsink or moving it around in the room.... not even the minimum noise. So if you use good cables (I've used twisted pairs double shielded low impedance Belden) and you twist also AC cables, EMI are surely not an issue.
Having said so, it's not a bad idea at all keeping PSUs in a separate case; but how will you bring 48V DC to the amp case?
If I had to build such an amp, only for simplicity, I would keep everything in a big MDF case (something around 60x40X15cm) with a pair of fans (one putting in, placed on the bottom outside the case, the other on the back, inside the case, putting air out) ,12mm diameter, slowed down to not produce any noise.
Just my 2c
Having said so, it's not a bad idea at all keeping PSUs in a separate case; but how will you bring 48V DC to the amp case?
If I had to build such an amp, only for simplicity, I would keep everything in a big MDF case (something around 60x40X15cm) with a pair of fans (one putting in, placed on the bottom outside the case, the other on the back, inside the case, putting air out) ,12mm diameter, slowed down to not produce any noise.
Just my 2c
Thanks for your 2c - Actually worth way more than that and some interesting thoughts.
The most surprising is the amps immunity to EMI. I would really think having a honkin big switching supply would somehow get some garbage into the amps. But wonderful if not!
Your idea of one large MDF case is also very good. I had thought I would need a metal case to help shield the amps, but if not I can build a wood or MDF case quite inexpensively. Will have to do some layouts on paper to see how much of this stuff I can fit in one large case.
Re the wiring if a separate case I thought I would use terminal strips rather than expensive plugs. Would have to work out the required wire sizes based on current draw. With individual supplies for each amp the wiring would be smaller of course, but more of it. Having amps and supplies in one case certainly is simpler!
Thanks also for your wiring and shielding tips.
Rod
MDF (or even "real" wood), it's better than metal, because does not reflect EMI inside the case and does not vibrate, keeping all more stable (see Faraday-Lenz law).
Of course these effects are quite not observable, but some people have listened to them...;-)
Thank you for your appreciation!
Of course these effects are quite not observable, but some people have listened to them...;-)
Thank you for your appreciation!
WOW - that is really touchy! Certainly argues for PSU not in separate case and and close to amp as possible.
Would be interesting to hear if others have had problems with this and how close they typically have the PSU to the amps.
Also might it help to twist the DC leads? I am assuming that you felt the problem was from the DC wires and not the AC ones, right?
Thanks for the tip.
Rod
short power leads?
I have a setup running class D boards with about 3 feet of leads. .. I have run two separate boards on it .. L25D and Iraud 350 The L25D boards ran normally if just one board was running .. with both boards running, they seemed to interfere with each other .. filtering the input signals helped some, but it appears according to what I can find that the 0 volt line is supposedly where the HF interference is playing the most. Unfortunately, can't test that because before I found that out, One of the L25D boards became useless, .. and one of the iraud 350 boards was basically DOA. I keep seeing stuff about how you are supposed to have separate power supplies especially the 0 volt lines to stop the boards from interfering with each other. I'm using an old pioneer sx1250 amplifier for the power supply ... has separate power supplies, but the 0 volt lines are connected (****!!!) I don't want to mess up the original amplifiers function, so I am thinking of ferrite beads or windings in the 0 volt line to choke out interference.
The L25D is weakly decoupled with 150 uf per pole ... the iraud 350 is strongly decoupled with almost 10k uf per pole .. The iraud 350 was supposed to drive 2 ohms, but I looked at the transistor specs it came with and didn't even want to try. I swapped the transistors out with IRFB4227 and amazingly, the board runs stable but I had to add heat sink to keep its temp down. its only because of the large heat sink on the 2092 chip that it can power the irfb4227. I don't like messing with the board though .. am going to replace the DOA iraud350 with iraud200 that uses irfb4227 standard. I am hoping that with strong decoupling, some 0 volt line filtering I can run the two together, and I think my chances are good because the 200 will run on a different frequency so interference will be less.
I have a setup running class D boards with about 3 feet of leads. .. I have run two separate boards on it .. L25D and Iraud 350 The L25D boards ran normally if just one board was running .. with both boards running, they seemed to interfere with each other .. filtering the input signals helped some, but it appears according to what I can find that the 0 volt line is supposedly where the HF interference is playing the most. Unfortunately, can't test that because before I found that out, One of the L25D boards became useless, .. and one of the iraud 350 boards was basically DOA. I keep seeing stuff about how you are supposed to have separate power supplies especially the 0 volt lines to stop the boards from interfering with each other. I'm using an old pioneer sx1250 amplifier for the power supply ... has separate power supplies, but the 0 volt lines are connected (****!!!) I don't want to mess up the original amplifiers function, so I am thinking of ferrite beads or windings in the 0 volt line to choke out interference.
The L25D is weakly decoupled with 150 uf per pole ... the iraud 350 is strongly decoupled with almost 10k uf per pole .. The iraud 350 was supposed to drive 2 ohms, but I looked at the transistor specs it came with and didn't even want to try. I swapped the transistors out with IRFB4227 and amazingly, the board runs stable but I had to add heat sink to keep its temp down. its only because of the large heat sink on the 2092 chip that it can power the irfb4227. I don't like messing with the board though .. am going to replace the DOA iraud350 with iraud200 that uses irfb4227 standard. I am hoping that with strong decoupling, some 0 volt line filtering I can run the two together, and I think my chances are good because the 200 will run on a different frequency so interference will be less.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.