Best charge pump for +/- 18v 100ma

Started by CheapPedalCollector, March 22, 2023, 07:55:31 PM

Previous topic - Next topic

CheapPedalCollector

I'm trying to decide what charge pump to use for +/- 18V supply from a single +18V input. I need 100ma for head room. I'm not finding anything from Ti that does more than 15 volts.

Would 7661 or 7662 work for this?

Jarno

Do you really need +/-18v, why is +/-15v not enough?
Would highly recommend the small Traco Power converters, not too expensive, don't need a lot of parts. But 100mA is already quite something, that's nearly 4watt max at +/-18v.

antonis

7661/7662 output current is restricted to 20mA..
(maybe a bit higher for voltage drop about 10%..)

And you'll need more than a single series-pass transistor for load drive current enhancement.. :icon_wink:

"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

CheapPedalCollector

Well yeah MXR EQ is +/- 18v and NE5532 (replacing 4558 with) has a max current draw of 16ma (*6) and I wanted to have some headroom, although I doubt the circuit draws that much current, but never know.

I really don't want to use a linear AC supply because of the hum, but I also want to use a quieted down version of the EQ I already own.

@antonis

Not really familiar with charge pumps in general, and that circuit doesn't look familiar. Is there some reading somewhere I can do?

Rob Strand

#4
QuoteWell yeah MXR EQ is +/- 18v and NE5532 (replacing 4558 with) has a max current draw of 16ma (*6) and I wanted to have some headroom, although I doubt the circuit draws that much current, but never know.
If you are really pressed for current you can run two charge pumps one for +18V and one for -18V.
However, if you start with 9V, under load you wont get 18V, it will be a bit less - that's life with charge pumps.  If you need to use diodes you need to use Schottky's to keep the drops down.

QuoteNot really familiar with charge pumps in general, and that circuit doesn't look familiar. Is there some reading somewhere I can do?
Datasheets and applications notes.  If you dig hard you can find quite a few applications notes.  IIRC Maxim had quite a few.

There's also roll-your-own charge pumps using NE555's.  Quite a few posts on this forum, a few recent ones on pushing the current as well.

Maybe this one?
https://www.diystompboxes.com/smfforum/index.php?topic=129366
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.


Kevin Mitchell

#6
Both the 4558 & 5532 have a max supply voltage of 30v. Bipolar 18v will surely cook em'.
I'd reconsider using an AC supply as it'll likely cost less in the end. Otherwise a DC-DC converter is the way.
  • SUPPORTER

antonis

Quote from: Kevin Mitchell on March 24, 2023, 01:02:51 PM
Both the 4558 & 5532 have a max supply voltage of 30v. Bipolar 18v will surely cook em'.

Both can be powered with 34V (+/-17) with no issue..
(but, IMHO, this should be done only if maximum headroom is required..) :icon_wink:
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

Kevin Mitchell

Quote from: antonis on March 24, 2023, 02:54:42 PM
Quote from: Kevin Mitchell on March 24, 2023, 01:02:51 PM
Both the 4558 & 5532 have a max supply voltage of 30v. Bipolar 18v will surely cook em'.

Both can be powered with 34V (+/-17) with no issue..
(but, IMHO, this should be done only if maximum headroom is required..) :icon_wink:
Ah you got me.
I guess max rating is only a suggestion and absolute max rating is what I should have looked for.
NE5532 is 44v and RC4558 is 36v
I suppose my next question would be if running them so hot would dramatically decrease their life?
  • SUPPORTER

antonis

Quote from: Kevin Mitchell on March 24, 2023, 03:36:44 PM
I suppose my next question would be if running them so hot would dramatically decrease their life?

Life is too short, anyway.. :icon_wink:

More seriously, both ICs are cheap enough to give them a try..

I've used 4558 (as well as 1458) at 38,6V (+33 / -5V6) and about 15mA output current for various benchtop power supplies without getting even warm..
(OK, maybe lukewarm..) :icon_redface:
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

CheapPedalCollector

Yeah 36v is the rating on the 4558, so I'm not worried about it. I'll check actually voltage levels in the pedal. This is a 70's one btw, not the modern Dunlop one.

@Rob yeah I want to use an 18v input so I only need to generate the negative voltage, I'll read that post. I did search around but didn't really find much that was related. It's pretty much a one off thing I'd like to get going.

@Jarno I bet that costs $10 or more, I'll check availability, don't think +/- 15v is what I need, but possibly.

Rob Strand

#11
You can see the positive doubler and negative doubler circuits here,
https://www.electronics-tutorials.ws/waveforms/555-circuits-part-2.html

It is possible to use a single NE555 and generate both positive and negative rails.
Lots of diodes and caps though.  Also if you generate both rails then the maximum
current will be half that of a single rail.  You can set the switch frequency yourself.

Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

amptramp

If you are looking for 100 mA, charge pumps are the wrong answer.  There are large losses from the switching FET's inside the charge pump IC and going to a separate oscillator (like a 555) and power FET's does not eliminate them.  It is best to use a flyback converter at this output level and they can be bought as complete units pretty cheap these days.  They would take less EMI filtering than a charge pump because the ramp upwards in on current is controlled by an inductor.  Shutoff of the inductor still causes noise but the methods of dealing with that are well known and not that difficult to implement.  The greater efficiency comes from the low current at turn-on, limited by the inductor.  FET losses are small when the current hasn't ramped up yet.

CheapPedalCollector

Yeah I don't think I can make the idea work. I don't want to use a flyback converter, such a thing is too big to fit in the enclosure.

I'll do some measurements on actual current draw from the circuit and see if the charge pump or 555 circuit is enough to handle it.