advertisement


Introducing the PowerReg

mikesnowdon said:
Cool.:) What kind of PSU would you advise with the powerreg?
Click to expand...

OK, I need to calculate this myself, so perhaps someone could correct any mistakes that may follow.

We need to make sure that the raw power supply voltage stays high enough to keep the regulator (LM317) happy. Let us say we want to generate 30V at the output. If we adjust R3 and R4, so that the gyrator bit drops an extra 4V over the LM317, then the output of the LM317 could be set to 34V. So the raw power supply must stay above roughly 34 + 3 = 37V.

The ripple voltage, due the raw power supply capacitor discharging, can be approximated using the formula Vripple ~ I / (2 x 50 x C) [I = current A, C = capacitance F, 50 is the mains frequency, and the 2 assumes full wave rectification]. So a current of 1A from a 10,000uF capacitor (0.01F) would drop the voltage by about 1V before recharging. Twice the current would double this, twice the capacitance would halve it. In other words, 1V per Amp.

So, for a 4A supply at 30V (120W), you would need a raw input voltage of 41V with a 10,000uF capacitor. Note that T3 would be dropping 11V, and so it will dissipate 44 Watts in heat, effectively "using up" 11/30 = 37% of the 4A*. We need to keep the voltage drop across T3 as small as possible, to avoid making too much toast. The bigger the capacitor, the less toast we need to make.

Obviously, the voltage drop caused by the transformer has been ignored in the above.

Andy

Edit: * because the input voltage is at 41V, and the output is 30V (37% higher), this allows the power dissipated by the output transistor to be ignored.
 
I'm using a 2x25V 160VA transformer with cheap 10000uF capacitors, it works great.

The output voltage of the PowerReg is set to 28V, I found it to be the optimal combination, sufficient dropout, and not too much heat. Works well with 4 and 8 ohm speakers.

Teddy
 
I tried PowerRegs in a LM3386 chip amp package from
Chipamp.com featuring a very good snubberized unregulated PS (designed by Carlos Filipe Machado) in which the ground/+ and ground/- are kept separated up to feeding the amp chip board. Now when I connected the Teddy PR's in between the PS and chimp amp board, its ground was starred and so invalidated the original design as it centred first the split incoming grounds into one point and then feeding the chimp split in two. The result is that I preferred the sound without the TeddPR's in place and now my question, am I right in blaming the starground in PR's as incompatible with this particular arrangement? Should I try any other grounding arrangement here?
Cheers wise men
 
I belive that is due to the dual bridge rectifiers on that particular PSU. You can run it from a center tapped trafo and use link wires to create a single bridge rectified PS where there is just one ground rail. Perhaps this setup would enable you to get the best from the PR?

On the other hand why not use 4 PR's? 2 for left +/- and 23 for right +/- ?
 
Cheers Mikesnowdon
I've been already powering the PS's from a dual centre tapped traffo (2x 25ish-0-25ish) from my bits box. I'll try to build a second Teddy PR board when I have spare time and, if you don't mind I'll PM you with the connection schematics to check I do it properly. But that will have to wait for a while.
Harthold
 
Hi.

If your using a 20-0-25v trafo then the dual bridge PSU wont work, unless you adapt it for centre tapped trafo as instructed in the manual, I assumne you did this though? Anyway if you want to go the dual bridge, seperate +/- gounds route then I'd go for seperate PRegs for left +/- and right +/- (assuming you have a 0-25, 0-25v non-CT trafo). Maybe Teddy agrees? ;)

Sure I'll do my best to help you but I think you should get in touch with Teddy really. He has way more knowledge and experience than me and should be able to elaborate on what I suggested. Im still a newbie. ;) Im sure when you get the best setup for implementing PReg's you amp will sound amazing. :)

Rgds,
Mike.
 
Setting R3 to 150K and R4 to 1M does not work so well for a 5V supply, because (with 10V from the LM317) you end up with only 1.3V across R3. A better solution may the following, which also avoids having R4 inside the first filter. Have you tried (rejected) this Teddy?

TR.jpg


I haven't built this yet, but I'd welcome any comments before I do. Thanks, Andy
 
Hi all, I want to ask question. Does anybody ever check using scope the volt signal noise output of TedReg ? I build this gyrator with Siklair output, and feeding to class A 2amp bias, at 13volt rails. I saw at output TedReg that audio signal is exist in the supply rails, seems output siklair not fast enough to counter this thing. And at input TedReg just after caps filter, that audio signal is not exist (just a little bit visible, very small).
While 50Hz noise only small in reduction.
Anyone has comments or same finding ? Maybe I have wrong wiring ?
 
That's not too surprising, there is no feedback mechanism in the teddyreg to reject load effects.

It's a passive filter, so while it is very effective at suppressing noise fed-through from the raw supply, it does have a finite, small output impedance. Therefore any load-related ripple current will create a load-related ripple voltage on the supply rail.

2A is quite a lot to draw even from this approach though; maybe your output device (sziklai) needs more current gain - or try a search here for Teddy's 'power reg'
 
With the PowerReg, driving a power amplifier at several amperes, the ripple at the PowerReg output is around 50-70mV, while without it, it can reach several volts...

It is important though to implement it correctly with the appropriate components, and the updates that I've published on this thread.

Teddy
 
Teddy.

Im interested in implementing the PowerReg in my CA 340a SE integrated. Its power amps are based on 1 x LM3886 p/ch and currently run on +/-34v . The trafo is 24-0-24v AC, about 120VA, 10'000uf smoothing caps. Im thinking that the basic config you used on you chipamp will be the best way to go. Can you provide a schematic for wiring or give me any advice on this? I have the 340a SE service manual which I could email to you.

Thanks in advance ;),

Mike.
 
I would be interested in an "idiots guide" to connecting a PowerReg to a chip amp and where and how in the power supply chain it should go. I assume Bridge Rect, smoothing caps and then PowerReg to Chipamp?
 
I have looked at the BOM for the PowerReg and it does not list part types for some positions. Can anyone advise a numpty for the following :-

U1 (+ve) LM317?
U2 (-ve) LM337?
Q3, Q4 & Q6

Thanks
 
"I assume Bridge Rect, smoothing caps and then PowerReg to Chipamp"
Correct

To which BOM are you referring?

Use the one in the first post. Note also that there were updates to the circuit posted later on this thread.
 
Teddy, I cannot see that the BOM at the beginning of this thread has the details for the components I mentioned in my last post?
 
The list is consistent with the diagram

R1 = 3.4K
R2 = 150R
R3 = 40-60K
R4 = not required
R5 = 40-60K
R6 = 100-250R

C1 = 0.1uF X7R Ceramic (optional)
C2 = 10-20uF Tantalum
C3 = 10-20uF Tantalum
C4 = 33uF Tantalum
C5 = 0.2-0.3uF X7R Ceramic
C6 = 10-20uF Tantalum/Oscon

T1 = SK117 GR
T2 = BC550C
T3 = D45H11
LR1 = LM317


There are no "U" or "Q" in this list nor in the diagram.
 
teddy_pardo said:
There are no "U" or "Q" in this list nor in the diagram.
Click to expand...
There are if Puffin is using the group buy board...

Puffin, are you building from scratch or using PowerReg boards from my group buy? If the latter (which I suspect to be the case) PM me with your email address and I will send the BOM.

Neil
 
You must log in or register to reply here.


advertisement


Back
Top