I built Pedja's LM338T based + and - regulator for my IGC (10k input, 220k feedback, 300VA per channel, pin 7 to gnd, 2*1000uF Panasonic FC caps, no input cap). After three weeks of breaking in the regulator I have the following to report:
The IGC sounds better WITHOUT the regulator! Some details:
1. The whole presentation has more air/ambience without the REG
2. High's are not as pronounced relative to mids without the REG
3. Highs are much more resolved without the REG-less sandy sounding!
4. Bass is not as tight without the REG, but transient information seems to come through more
5. But overall, my IGC is making music again and not sounding so hi-fi-ish!
Even my wife seems happier with the sound...
Anyone tried the solid state (Pedja) buffer yet? What about the tube based buffer?
Now time to try Thorsten's ULTIMATE power supply.
Ryan
The IGC sounds better WITHOUT the regulator! Some details:
1. The whole presentation has more air/ambience without the REG
2. High's are not as pronounced relative to mids without the REG
3. Highs are much more resolved without the REG-less sandy sounding!
4. Bass is not as tight without the REG, but transient information seems to come through more
5. But overall, my IGC is making music again and not sounding so hi-fi-ish!
Even my wife seems happier with the sound...
Anyone tried the solid state (Pedja) buffer yet? What about the tube based buffer?
Now time to try Thorsten's ULTIMATE power supply.
Ryan
What do you do if you want to regulate more than 24V or 1.5A? Can you parrellel regulators to bump up the amperage rating, like two 24V/1.5A regulators parrelleled to give like a 24V/~3.0A? Also what if you want 35V regulated? Even possible? Or do most places just not list anything greater than 24V/1.5A cause there not as popular.
Responses
TBLA,
Have tried 50uF*2 at the output of the LM338 and the bass went lower, was more expansive, but seemed to overwhelm the music. This was while still using 2*1000uF on the LM3875, i.e. 1000 on + and 1000 on -
Removal of the 50uF*2 at the output cured the "mighty weighty bass" problem.
Removal of the reg cured the harsh hi problem as described, but most importantly bought music back.
Hybrid fourdoor
The use of LM338T allows for 5A as opposed to the traditional 1-1.5A of the KM317. To use LM338 for a - supply, build two positive supplies using LM338, but then connect the OUTPUT of the second LM338 to GROUND. Voltages are relative to ground, so the second LM338 effectively provides a 5A regulated supply through its ground voltage relative to the system ground.
TBLA,
Have tried 50uF*2 at the output of the LM338 and the bass went lower, was more expansive, but seemed to overwhelm the music. This was while still using 2*1000uF on the LM3875, i.e. 1000 on + and 1000 on -
Removal of the 50uF*2 at the output cured the "mighty weighty bass" problem.
Removal of the reg cured the harsh hi problem as described, but most importantly bought music back.
Hybrid fourdoor
The use of LM338T allows for 5A as opposed to the traditional 1-1.5A of the KM317. To use LM338 for a - supply, build two positive supplies using LM338, but then connect the OUTPUT of the second LM338 to GROUND. Voltages are relative to ground, so the second LM338 effectively provides a 5A regulated supply through its ground voltage relative to the system ground.
Dr.H
It looks to me like your regulators are WAY to small for the chip amp you are building. If you have a current probe for your o-scope you will see that current peaks will limit with only 1.5 amp reg. chips. With out the regulators if you check the current of the + and - rails you will see peaks that go over the 1.5 amp limit of the regulators. That is why the regulators you used make the amp sound that way.
Now of you were to make power supplies that could deliver 10 or 20 amps of peak current at the voltage required I am sure that you conclusion might be different. No hum or noise as related to power supply, very wide bandwidth and so stable that any load is not a problem.
I have been using BIG regulated power supplies with all of my power amps. for years and always will.
I know I'll take a few shots for this view but I have tried it many ways (big caps, high dollar caps, bypass caps, CLC filters, LC filters) but the way that always works is a big regulator that is done as well as the amp it powers, end of the story for me. After all the power supply is in series with the amp as far as the DC goes right?
Later BZ
OK I see you posted as I was typing that you used 5 amp regulators that would help but could still be a limiting factor. Tell me when you take out the regulators what is you rail voltage? Do you some how reduce it to the same voltage as with the regulators?
It looks to me like your regulators are WAY to small for the chip amp you are building. If you have a current probe for your o-scope you will see that current peaks will limit with only 1.5 amp reg. chips. With out the regulators if you check the current of the + and - rails you will see peaks that go over the 1.5 amp limit of the regulators. That is why the regulators you used make the amp sound that way.
Now of you were to make power supplies that could deliver 10 or 20 amps of peak current at the voltage required I am sure that you conclusion might be different. No hum or noise as related to power supply, very wide bandwidth and so stable that any load is not a problem.
I have been using BIG regulated power supplies with all of my power amps. for years and always will.
I know I'll take a few shots for this view but I have tried it many ways (big caps, high dollar caps, bypass caps, CLC filters, LC filters) but the way that always works is a big regulator that is done as well as the amp it powers, end of the story for me. After all the power supply is in series with the amp as far as the DC goes right?
Later BZ
OK I see you posted as I was typing that you used 5 amp regulators that would help but could still be a limiting factor. Tell me when you take out the regulators what is you rail voltage? Do you some how reduce it to the same voltage as with the regulators?
Since I knew Ryan offline from private emailing during the time he was building some stuff suggested at my site, I know also that he, else than the fact he had an objections about the depth of the bass, has positive initial findings (20 days ago) about this supply. That was the last info I had from him.
Of course, during the time, positive initial findings might appear as false ones, and to make this very clear, I in fact consider the long term listening as the best evaluating method. Also, I highly value someone’s readiness, after some time spent for building, to give up of that as a wrong thing.
However, as someone who invested some time and effort to help the man that reported positive initial findings, I feel very surprised to see this PS regulator bad for IGC sound announcement. Not because he changed his mind but because of the way I meet this fact.
The rest is perfectly fine with me. Of course, including the announcing of the findings and conclusions I obviously do not agree.
Pedja
Ps: To be very clear about this, I recommend the regulated supply as I did before. Those suspicious about the LM338’s ability to drive the LM3875 chip should check this page.
Of course, during the time, positive initial findings might appear as false ones, and to make this very clear, I in fact consider the long term listening as the best evaluating method. Also, I highly value someone’s readiness, after some time spent for building, to give up of that as a wrong thing.
However, as someone who invested some time and effort to help the man that reported positive initial findings, I feel very surprised to see this PS regulator bad for IGC sound announcement. Not because he changed his mind but because of the way I meet this fact.
The rest is perfectly fine with me. Of course, including the announcing of the findings and conclusions I obviously do not agree.
Pedja
Ps: To be very clear about this, I recommend the regulated supply as I did before. Those suspicious about the LM338’s ability to drive the LM3875 chip should check this page.
Would adding reserve caps after the reg improve the situation a bit w/ peak current? I see no real reason to not add more than the 470uf mentioned, as long as they are bypassed.
Also, w/ a seperate regulated PS, if you are using a long umbilical I've read that the inductance of the wires has been known to cause problems. The mentioned solution to it was to add an additional pair of 5uf caps at the amp end of the cord(there were 5uf film caps at the power end in the design).
Also, w/ a seperate regulated PS, if you are using a long umbilical I've read that the inductance of the wires has been known to cause problems. The mentioned solution to it was to add an additional pair of 5uf caps at the amp end of the cord(there were 5uf film caps at the power end in the design).
Try a discrete regulator, a simple zener/emitter follower will work well. Use a big power BJT and it'll output lots of current.
I'm not saying the LM338 won't work well, I never tried it. I'm just recomending this method, no current limiting mind you, but I've always found that simple emitter follower regs sound very good.
I'm not saying the LM338 won't work well, I never tried it. I'm just recomending this method, no current limiting mind you, but I've always found that simple emitter follower regs sound very good.
Some responses
Thanks for all the responses.
To provide some context, the rest of my system is:
Proceed PDT3 transport
Theta Probasic II
IGC
Royd doublet speakers (89db)
TBLA,
I think it pretty hard to make mistakes with a simple 3 terminal regulator. I am interested in your implementation. Mine is as follows:
MUR860 diodes*4, 4700uF*2 (generic), LM338T, setup for 22V, 50uF (Panasonic FC) from ADJ to ground. This then feeds the LM3875 which has 1000uF at the pins (each PS pin 1000uF). Previously I used 50uF*2 at the output of the LM338, but found bass a bit overwhelming.
Pedja,
Seems we have a difference of opinion about the impact of the regulator, but I am open to further discussion and sharing of opinions. I am sure that many other will benefit from teh sharing of these views on DIYAUDIO.COM
Others:
Yes, maybe trying a different regulator would be useful, however what I REALLY want to know is HAVE YOU BUILT A REGULATOR FOR THE IGC AND HOW HAS IT AFFECTED YOUR OPINION OF THE SOUND?
Thanks for discussion.
Ryan
Thanks for all the responses.
To provide some context, the rest of my system is:
Proceed PDT3 transport
Theta Probasic II
IGC
Royd doublet speakers (89db)
TBLA,
I think it pretty hard to make mistakes with a simple 3 terminal regulator. I am interested in your implementation. Mine is as follows:
MUR860 diodes*4, 4700uF*2 (generic), LM338T, setup for 22V, 50uF (Panasonic FC) from ADJ to ground. This then feeds the LM3875 which has 1000uF at the pins (each PS pin 1000uF). Previously I used 50uF*2 at the output of the LM338, but found bass a bit overwhelming.
Pedja,
Seems we have a difference of opinion about the impact of the regulator, but I am open to further discussion and sharing of opinions. I am sure that many other will benefit from teh sharing of these views on DIYAUDIO.COM
Others:
Yes, maybe trying a different regulator would be useful, however what I REALLY want to know is HAVE YOU BUILT A REGULATOR FOR THE IGC AND HOW HAS IT AFFECTED YOUR OPINION OF THE SOUND?
Thanks for discussion.
Ryan
maybe the problem isn't with the current limit per se, but the regulator's ability to supply that current quickly. For example, when the amp needs to go from 100ma output to 2A output in 1ms, and the regulator just couldn't react fast enough to supply the current that quickly.
Those chips have pretty high psrr so I really doubt how much improvement one can get from a regulated power supply.
Those chips have pretty high psrr so I really doubt how much improvement one can get from a regulated power supply.
Pedja said:
there may be a difference between the PS for a class-a amp and that for a class-b amp. in a class-a amp, the current draw from the PS is abot the same, and it fluctuates widely for a class b amp.
so if the regulator does have problem reacting to fast current swings, it will work fine in a class-a amp but not in a class-b amp.
hi,
it will be a mistake if you don't put a cap right on the lt1083cp regs output.....i use 1000uF and then there is some distance to the lm chip - as short as possible - and there you use another cap (470uF elna silmic II, 63 volts) perhaps bypassed with a film (wima fkp3) yo
edit, reg input is elna lpo 4700uF and a film right at the inputpin - and adjustpin resistor is bypassed to ground with 100uF or even more......
Since I have one channel unregualted and one regulated (for comparison purposes), I will try your "1000uF at output of LM338" idea. In my current set-up there is about 10cm between the LM338 output pin and the IGC power pins. PS and IGC are in the same box. Note that adding the additional 1000uF cap will of course result in a total of 2000uF per IGC PS pin. But, I'll try it and let you know.
In response to a previous post, the reg was set for 22V, but without the reg, I am running the IGC at 31V. I have read that the higher the unregulated voltage, the tighter the bass, although lower unreg voltage gives lower bass.
Ryan
In response to a previous post, the reg was set for 22V, but without the reg, I am running the IGC at 31V. I have read that the higher the unregulated voltage, the tighter the bass, although lower unreg voltage gives lower bass.
Ryan
Squalish said:
millwood said:
millwood said:
From National LM138/338 datasheet:
A unique feature of the LM138 family is time-dependent current
limiting. The current limit circuitry allows peak currents
of up to 12A to be drawn from the regulator for short periods
of time. This allows the LM138 to be used with heavy transient
loads and speeds start-up under full-load conditions.
Under sustained loading conditions, the current limit decreases
to a safe value protecting the regulator.
Pedja
About the output capacitors using LM338, National datasheet again:
Although the LM138 is stable with no output capacitors, like
any feedback circuit, certain values of external capacitance
can cause excessive ringing. This occurs with values between
500 pF and 5000 pF. A 1 µF solid tantalum (or 25 µF
aluminum electrolytic) on the output swamps this effect and
insures stability.
I don't use output caps other than those mounted at the chips.
Situation with LT1083 is a bit different, at least according to LT's datasheet:
The circuit design used in the LT1083 family requires the
use of an output capacitor as part of the device frequency
compensation. For all operating conditions, the addition of
150uF aluminium electrolytic or a 22uF solid tantalum on
the output will ensure stability. Normally, capacitors much
smaller than this can be used with the LT1083. Many
different types of capacitors with widely varying characteristics
are available. These capacitors differ in capacitor
tolerance (sometimes ranging up to ±100%), equivalent
series resistance, and capacitance temperature coefficient.
The 150uF or 22uF values given will ensure stability.
Pedja
Although the LM138 is stable with no output capacitors, like
any feedback circuit, certain values of external capacitance
can cause excessive ringing. This occurs with values between
500 pF and 5000 pF. A 1 µF solid tantalum (or 25 µF
aluminum electrolytic) on the output swamps this effect and
insures stability.
I don't use output caps other than those mounted at the chips.
Situation with LT1083 is a bit different, at least according to LT's datasheet:
The circuit design used in the LT1083 family requires the
use of an output capacitor as part of the device frequency
compensation. For all operating conditions, the addition of
150uF aluminium electrolytic or a 22uF solid tantalum on
the output will ensure stability. Normally, capacitors much
smaller than this can be used with the LT1083. Many
different types of capacitors with widely varying characteristics
are available. These capacitors differ in capacitor
tolerance (sometimes ranging up to ±100%), equivalent
series resistance, and capacitance temperature coefficient.
The 150uF or 22uF values given will ensure stability.
Pedja
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.