| tomchaoda |
I am about to place the Inductors (2x2.2 mH) in my Aleph 5 to form a pi filter. Few simple questions (as I am using this type the first time):
1. How should I place the Inductors (flat or raise up? Should their axle be parallel or perpendicular is fine?). I ask this, because my case is limited in space.
2. Should I isolate my Inductors to the chassis (I think I should, but is it requried?)
Thanks
Thomas |
|
|
| roddyama |
| You can mount them either way as long as you keep them away from the low level signals. They should be isolated from chassis. |
|
|
| Tyimo |
Hi!
Another PSU question: need I use fuse between PSU and Amp board or it is not important? What kind of type and how big? ( slow blow, 4A? )
Thanks ! |
|
|
| roddyama |
| This is a debatable subject. Myself, I wouldn't put a fuse in the DC line. I think it would become a noise source. There are those that have used fuses or breakers, and even relays, as I recall. The fuse size would be at least 3A per rail and 4A should be fine. |
|
|
| Tyimo |
Thanks!
And what about with Main AC filter? Is it good or not? |
|
|
| roddyama |
| IMO anything that can clean up the input AC would help the amp to sound better. Although you may have a problem finding (or paying for) a line filter that will take the continuous 300W draw of an Aleph5 and live through the inrush current that will go to charging the PSU caps on power-up. |
|
|
| dantwomey |
| quote: | Originally posted by roddyama
IMO anything that can clean up the input AC would help the amp to sound better. Although you may have a problem finding (or paying for) a line filter that will take the continuous 300W draw of an Aleph5 and live through the inrush current that will go to charging the PSU caps on power-up. |
When the topic of line filters comes up I've noticed lots of referrals to Jon Risch's DIY Line Filter.
Regards,
Dan |
|
|
| tomchaoda |
If I wanted to test the filtering effect of either CLC or CRC, what Resistor value should I use? How can I calculate it?
One example I saw is using DALE 0.5R / 50W and another one I saw is 40R / 40W. This is a big difference!
Can anyone shed some light on this?
Thanks
Thomas |
|
|
| Blues |
| Depends on how much you'd like your final rail values would be. 0.5R would be less voltage drop compared to 40R. I think the effect of an R in a CRC filter is the increase in the discharge time of the capacitor, thereby decreasing the ripple. The L meanwhile in a CLC has 2 effects on the ripple with its dcR doing what an R does and inductance filling in the valley between the cap discharge and the next peak of the bridge signal. The inductor's degree of lag from the applied voltage depends on the value of the dcR. |
|
|
| SteveA |
The Aleph 5 runs on +/- 34 v (give or take). I have a 30-0-30 transformer which without any CLC or CRC added would give about 42 v rails. I understand adding CRC to the PS will reduce the voltage. Question: Would it be possible to use this transformer with the Aleph 5 with suitable CLC? Or, said differently, what inductor could I choose which would result in acceptable rail voltages?
Thanks,
SteveA |
|
|
| roddyama |
| quote: | Originally posted by SteveA
The Aleph 5 runs on +/- 34 v (give or take). I have a 30-0-30 transformer which without any CLC or CRC added would give about 42 v rails. I understand adding CRC to the PS will reduce the voltage. Question: Would it be possible to use this transformer with the Aleph 5 with suitable CLC? Or, said differently, what inductor could I choose which would result in acceptable rail voltages?
Thanks,
SteveA |
Steve,
Theoretically your transformer will give you 42V rails. Actually it will be closer to +/-40V, low load, give or take a volt or so. You have to subtract the drop across the bridge and the secondary winding (depending on the transformer design). Than you subtract for the DCR of the L in the CLC. On top of that (or factored in), as the current draw increases, all of these voltage drops increase. You will probably end up with +-37V rails which would work fine as long as the other components can take the voltage. |
|
|
| Petter |
| quote: | Originally posted by SteveA
The Aleph 5 runs on +/- 34 v (give or take). I have a 30-0-30 transformer which without any CLC or CRC added would give about 42 v rails. I understand adding CRC to the PS will reduce the voltage. Question: Would it be possible to use this transformer with the Aleph 5 with suitable CLC? Or, said differently, what inductor could I choose which would result in acceptable rail voltages?
Thanks,
SteveA |
Search for "Duncan PSUD2" on the web for very simple PSU software with great utility.
You need to use an LC filter. This will signficantly lower the output voltage.
Petter |
|
|
| SteveA |
Petter-
Thanks!!
STeveA |
|
|
| SteveA |
Petter-
I am trying to understand how to use PSUD2. For example, I am trying to model the Aleph 5 psu from Kristijan's site as a "tutorial":
http://web.vip.hr/pcb-design.vip/power-5.html
But I can't seem to make the design look like the supply which Kristijan has depicted. If I could see the step by step process in modeling a PSU in this program it would be helpful.
Thanks,
Steve |
|
|
| roddyama |
| quote: | Originally posted by SteveA
Petter-
I am trying to understand how to use PSUD2. For example, I am trying to model the Aleph 5 psu from Kristijan's site as a "tutorial":
http://web.vip.hr/pcb-design.vip/power-5.html
But I can't seem to make the design look like the supply which Kristijan has depicted. If I could see the step by step process in modeling a PSU in this program it would be helpful.
Thanks,
Steve |
Hi Steve,
You won't be able to model parallel capacitors in Duncan's PSU program. You have to add them together as a single capacitance. Kristijan's PSU modeled in PSUD2 will consist of:
Trans => Bridge => 44kuf => 2.2mH => 44kuf => Load (or current draw) |
|
|
| SteveA |
Hi Rodd-
My problem is (I did find the part you made about parallel caps) that I can only make the program do the C or the LC portion. I haven't figured out how to append the LC to the C.
I have the "parts" (except for the inductors) for the PSU so I suppose I could assemble it and see how it measures. I would like to know what the final voltage would be without having to order an assortment of inductors do the trial and error thing.
That fact (and the fact that I am trying to learn electronics as I go; chemist by training) led to my initial question.
Thanks,
Steve |
|
|
| SteveA |
Rodd-
I played around a bit and figured out how to do the CLC in the design.
Steve |
|
|
| roddyama |
| quote: | Originally posted by SteveA
Rodd-
I played around a bit and figured out how to do the CLC in the design.
Steve |
I knew you could do it.:up:;) |
|
|
| Tyimo |
Hi!
Some more question:
-Should I use 2 pieces 22R 3W thermistors paralel instead of the 1 10R5W-er?
-Should I put them between AC fuse and switch?
-How many Amps draw the stereo Aleph5 from the AC socket?
-What is the best PSU-caps arrangement-conection?
Thanks! |
|
|
| mpopovics |
Hello Tyimo!
1. You can use 2 paralel, but I recommend the 5ohm type
"...TH1 is a power thermistor used to connect the circuit and chassis ground to the AC outlet
ground. It will normally operate at 5 ohms, suppressing ground loops in the system, but will
drop to a low impedance if significant current is passed through it...."
2. Yes, you can put between AC fuse and switch
3. From service manual:
"...The voltage rails of the supply are at 34 volts, and each channel draws approximately 150
watts.
...
F1 is a slow blow fuse, set at 4 amps for 100-120 volt operation, and 2 amps for 220-240
volt operation."
4. In my opinion CLC, check out kristijan's homepage (link above)
BR, |
|
|
| roddyama |
| quote: | Originally posted by Tyimo
-Should I use 2 pieces 22R 3W thermistors paralel instead of the 1 10R5W-er? |
If you are going to use them for current limiting, you should take a look at this thread about the thermistor that Pass Labs uses. Assuming you have 220VAC, you might think about 2 - 10ohm in series to get the inrush current down.| quote: | Originally posted by Tyimo
-Should I put them between AC fuse and switch? |
I think either way will work.| quote: | Originally posted by Tyimo
-How many Amps draw the stereo Aleph5 from the AC socket? |
The manual says 300W, so it's 300/VAC = Amps.
-What is the best PSU-caps arrangement-conection? [/B][/QUOTE]I like the CLC, but most any other arrangments will work. |
|
|
| Tyimo |
Mpopovics and Roddyama!
Thanks!
| quote: | | "F1 is a slow blow fuse, set at 4 amps for 100-120 volt operation, and 2 amps for 220-240 |
So, could I use a 4A 250VAC main AC filter for the stereo Aleph5? Isn't too weak?
Could be only 1 piece 3W-er thermistor is enough for current limiting?
Thanks |
|
|
| mpopovics |
2 amps is just fine for stereo
P=U*I -> I=P/U=300W/240V=1,25A
BR, |
|
|
| SteveA |
I have seen 2.2 mH on Kristijan's site but I believe I have seen recommendations of 1.1 also. A recommendation from someone in the know would be appreciated.
SteveA |
|
|
| roddyama |
| quote: | Originally posted by SteveA
I have seen 2.2 mH on Kristijan's site but I believe I have seen recommendations of 1.1 also. A recommendation from someone in the know would be appreciated.
SteveA |
According to Nelson Pass. |
|
|
| SteveA |
Hi Rodd-
Thanks. In the reference was a comment that most of the capacitance should be after the inductor. Is that mandatory? I have some 50v 68000 caps which I was planning on using two per channel with the inductor between them. Duncan's PSU indicates a voltage of +/- 31.9 v. This would seem to be in the hunt (yes?).
I also have some 75 v 22000 caps but they're big as Buicks.
Steve |
|
|
| roddyama |
| quote: | Originally posted by SteveA
Hi Rodd-
Thanks. In the reference was a comment that most of the capacitance should be after the inductor. Is that mandatory? I have some 50v 68000 caps which I was planning on using two per channel with the inductor between them. Duncan's PSU indicates a voltage of +/- 31.9 v. This would seem to be in the hunt (yes?).
I also have some 75 v 22000 caps but they're big as Buicks.
Steve |
I know the 50X68kuf. Those are pretty good caps. There are a few recent projects using them. I'll probably use them myself in my up coming amp. |
|
|
| Nelson Pass |
| quote: | Originally posted by SteveA
In the reference was a comment that most of the capacitance should be after the inductor. Is that mandatory |
No, but as a practical matter, we usually get a batch of the
same kinds of caps, so we tend to put the same capacitance
before the inductor as after. The only danger is in putting
way too little on either side, but once you have at least, say,
10,000 uF before and after, you're on your way. You have to
insure that the "before" caps meet the ripple current rating,
and after that, it's pretty arbitrary, as increasing capacitance
on either side reduces the noise that the circuit sees by
inverse proportion.
Having too much capacitance on either side can be a problem
with regard to initial turn-on inrush currents, but there are
plenty of ways to deal with that, thermistors being one popular
approach.
:cool: |
|
|
| SteveA |
Thanks Mr Pass-
I will be using a rectifier which I saw Wayne Sankey use on his Aleph 2 project. The board has provisions for two thermistors. I hope that will be sufficient.
SteveA |
|
|
| Sandy H. |
I am interested in this thread as I hope to begin the PS on an Aleph project soon as well.
When modeling the PS in PSUD2, trying both a CRC and a CLC arrangement, the CRC seems to have less ripple. If this is correct, then is the only reason a CLC is more often recommended (it seems) because of the lower power loss? If I am doing the simulation correctly, then it seems that the difference between a 2.2mH and a 4.4mH inductor would be around 1/2 a volt. I was always under the impression that it was more.
Are the above observations consistent with the group's experiences? My real concern is that I too am hoping to use the 68000uf 50V caps, but was planning on an Aleph 2. It seems that I’m on the edge if I use a standard Avel 40-0-40, which modeled 44.5VRMS at the rails with a 7.5A (est 300W/40V) constant current draw. However, the peak on C1 is 49.8V, which seems a bit too close for comfort with my knowledge and the fact that they are 50V surplus caps. Should I consider getting some 63V caps for the first set of caps or just play it conservative and go for a 35 or 37V torroid? Finally, with low/no load conditions, the caps are way over their limits (around 57V). Is this an issue, since it is my understanding that a class A amp always draws constant current due to the constant dissipation across the transistors (the bias current) regardless of signal? It seems that the only time I could get in trouble would be during the trimming of the amp during the initial testing.
The above is how I understand these items from all the reading here about amp design. Please point out anything I have misunderstood, as I don’t want to build on a false foundation.
Sandy. |
|
|
| SteveA |
I assembled the PSU (for Aleph 5). I used a 30-0-30 toroid with a 2 mH inductor "sandwiched" between two 68000 50 v caps (for each rail). Duncan's software suggested I should see +/- 32v. I measure +/- 43. I used the rectifier board (and diodes) from the Aleph 2 posted on Passdiy by Wayne Sankey. Why the difference between the predicted and observed?
Sure wish this was chemistry.
Steve |
|
|
| roddyama |
| quote: | Originally posted by Sandy H.
I am interested in this thread as I hope to begin the PS on an Aleph project soon as well.
When modeling the PS in PSUD2, trying both a CRC and a CLC arrangement, the CRC seems to have less ripple. If this is correct, then is the only reason a CLC is more often recommended (it seems) because of the lower power loss? If I am doing the simulation correctly, then it seems that the difference between a 2.2mH and a 4.4mH inductor would be around 1/2 a volt. I was always under the impression that it was more.
Are the above observations consistent with the group's experiences? My real concern is that I too am hoping to use the 68000uf 50V caps, but was planning on an Aleph 2. It seems that I’m on the edge if I use a standard Avel 40-0-40, which modeled 44.5VRMS at the rails with a 7.5A (est 300W/40V) constant current draw. However, the peak on C1 is 49.8V, which seems a bit too close for comfort with my knowledge and the fact that they are 50V surplus caps. Should I consider getting some 63V caps for the first set of caps or just play it conservative and go for a 35 or 37V torroid? Finally, with low/no load conditions, the caps are way over their limits (around 57V). Is this an issue, since it is my understanding that a class A amp always draws constant current due to the constant dissipation across the transistors (the bias current) regardless of signal? It seems that the only time I could get in trouble would be during the trimming of the amp during the initial testing.
The above is how I understand these items from all the reading here about amp design. Please point out anything I have misunderstood, as I don’t want to build on a false foundation.
Sandy. |
| quote: | Originally posted by SteveA
I assembled the PSU (for Aleph 5). I used a 30-0-30 toroid with a 2 mH inductor "sandwiched" between two 68000 50 v caps (for each rail). Duncan's software suggested I should see +/- 32v. I measure +/- 43. I used the rectifier board (and diodes) from the Aleph 2 posted on Passdiy by Wayne Sankey. Why the difference between the predicted and observed?
Sure wish this was chemistry.
Steve |
Sandy, Steve,
Duncans is good for a quick look to get you in the ballpark. You have to have the DCR of the inductors and the ESR of the caps set as real as possible. Same holds for the other components. Then you need to very the load as it will vary in real life. In the case of the Aleph amps where there's a large current draw, you can get away with marginal cap voltage values. However, if you use tricks like limiting the bias current during non-use, or if you use a safety disconnect on the DC rails where the load current is allowed to go to a low value or zero, the cap voltage will go to the maximum no load voltage. This value can be found by setting the load to 0A in PSU2. (Sandy, applying voltage with no-load during testing and setup is also a place where you can over voltage the caps.)
Steve,
You should see voltages in the 40+ Volt range if you're measuring the static supply. The voltage drops come in to play when you have a load current flowing. You need to attach a load resistor on the supply output to get a true picture of the "in use" Voltages. (Be sure the load resistor is of a high enough wattage.) |
|
|
| SteveA |
Hi Rodd-
Thanks, I inserted a 10R 22W resistor between the + and - outputs and measured ~33.5 volts. However, within seconds the resistor was starting to smoke. Is 22W not adequate?! Is it too much? I thought I would simulate an 8 ohm load but all I had available was a 10 ohm 22 Watt resistor. What should I have used?
Even though the output side of the supply seems to be providing the correct voltage I am concerned about frying the circuit boards were they to be connected at this stage.
Thanks again
Steve |
|
|
| Blues |
| SteveA, you were dissipating 115W on a 22W resistor...:devilr: |
|
|
| roddyama |
Steve,
The Aleph amp output sees a different load than its PSU. If you're testing the actual PSU alone you need to consider the load it sees. The total power consumption of the Aleph 5 is about 300W. Some of that will be consumed by the supply itself, but for testing the PSU, it's a good place to start.
300W delivered by 2 X 35V = 70V mean you need a total load of a little over 4 amps. That would equate to a resistor of 70V/4A ~ 18ohms @ 300W. For this you can use 2 - 10ohm resistors @ 150W, with the resistors tied to the + rail to ground and ground to the - rail.
First, you should be using a variac so you can bring the voltage up slowly. You can increase the resistance to simulate a lighter load (~2A at idle at the PSU output). |
|
|
|
