|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
28th Oct 2020, 4:30 pm | #1 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
SMPSU + Linear regulator combo?
Could I use a linear regulator like a LM317 to drop the voltage off a SMPSU? Good idea or a nightmare on a stick? I'm thinking something like a 12v SMPSU into a big cap, then into a LM317.
Why? Saves space and a tfmr. Andy.
__________________
Curiosity hasn't killed this cat...so far. |
28th Oct 2020, 4:37 pm | #2 |
Hexode
Join Date: Mar 2014
Location: Fareham, Hampshire, UK.
Posts: 482
|
Re: SMPSU + Linear regulator combo?
Should work, don’t forget two 100 nF caps to input and output and maybe two larger caps as well, 10 uF or so to prevent oscillation.
__________________
Failing to prepare is preparing to fail ... |
28th Oct 2020, 4:37 pm | #3 |
Moderator
Join Date: Jun 2003
Location: Oxford, UK
Posts: 27,787
|
Re: SMPSU + Linear regulator combo?
Why wouldn't it work? DC is DC, wherever it comes from.
Use a protection diode too if you're being careful, so that there's no opportunity for a cap to discharge 'backwards' into the regulator. |
28th Oct 2020, 4:47 pm | #4 |
Moderator
Join Date: Dec 2010
Location: Exeter, Devon and Poole, Dorset UK.
Posts: 6,823
|
Re: SMPSU + Linear regulator combo?
Hi Andy whats the big cap for?
SMPS usually have good regulation an respond pretty well to step changes in current. Cheers Mike T
__________________
Invisible airwaves crackle with life or at least they used to Mike T BVWS member. www.cossor.co.uk |
28th Oct 2020, 4:59 pm | #5 |
Heptode
Join Date: Feb 2006
Location: Marlborough, Wiltshire, UK.
Posts: 915
|
Re: SMPSU + Linear regulator combo?
This was (maybe still is) a common technique when switched mode psu were less developed and/or there was a large I/O difference. The smps would do the “rough” work and provide just enough voltage for the linear to give good performance without too much heat.
|
28th Oct 2020, 5:01 pm | #6 |
Hexode
Join Date: Apr 2007
Location: W Yorks, UK.
Posts: 406
|
Re: SMPSU + Linear regulator combo?
|
28th Oct 2020, 5:04 pm | #7 |
Hexode
Join Date: Mar 2014
Location: Fareham, Hampshire, UK.
Posts: 482
|
Re: SMPSU + Linear regulator combo?
Good point, thanks Paul.
__________________
Failing to prepare is preparing to fail ... |
28th Oct 2020, 5:14 pm | #8 | |
Guest
Posts: n/a
|
Re: SMPSU + Linear regulator combo?
Quote:
|
|
28th Oct 2020, 5:24 pm | #9 |
Dekatron
Join Date: Jun 2009
Location: Staffordshire Moorlands, UK.
Posts: 5,263
|
Re: SMPSU + Linear regulator combo?
some industrial smpsu's with multiple output voltages use IC regulators like this, built in already from the main output.
__________________
Kevin |
28th Oct 2020, 8:02 pm | #10 | |
Octode
Join Date: Feb 2009
Location: Bletchley, Buckinghamshire, UK.
Posts: 1,205
|
Re: SMPSU + Linear regulator combo?
Quote:
Wanting to get away from their traditional (and very heavy) toroidal transformer supplies, the manufacturer had decided to use bought-in SMPS modules followed by their own linear regulators. I think they got through three manufacturers before they found one that was reliable enough to avoid wrecking their reputation for a world-class product. Because a failed mixer PSU would be a show-stopper, it was standard practice to run with two supplies, one being redundant, with diode combiners. But in this case we always made sure there were three... plus another spare stashed away in the flightcase of contingency (GOOS) kit. Happily, SMPSU's are now a mature technology and normally reliable. |
|
29th Oct 2020, 7:14 am | #11 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: SMPSU + Linear regulator combo?
Thanks chaps, I was worrying for nothing. I was wondering if the "noise" off the SMPSU might interfere with the 317 in some way, so thought a big cap might be needed before the 317, then wondered if SMPSU's might find a capacitive load hard to drive. I have seen ATX supplies with 78xx after regs, hence the idea.
I'll, of coarse be sure to include diodes, caps etc as per datasheet. Andy.
__________________
Curiosity hasn't killed this cat...so far. |
29th Oct 2020, 4:17 pm | #12 |
Octode
Join Date: Nov 2013
Location: Penrith, Cumbria, UK
Posts: 1,993
|
Re: SMPSU + Linear regulator combo?
I tried using a SMPS wallwart to power a small ClassD amp module and an associated chip based tone and preamp module.
It was very disappointing in that there was a lot of noise and hum introduced into the circuit. I tried a 317 based reg board on it in case it wasnt properly regulated and it made no difference. Some SMPS just arent up to the job if you need them for low noise audio or other critical applications. I dont really understand how they work but i was told it had something to do with noise injected on the earth/0V rail. I may have completely misunderstood that part, I find SMPS to be black magic and unfathomable and they have no place in any of my analogue amateur Audio or radio builds. I imagine theres some professional members here who know the ins and outs and ways to filter the noise out. It must be possible as my Squeezebox touch relies on a SMPS. Like you I was thinking how nice to have a use for a bunch of spare SMPS wallwarts. I seem to have amassed a set of BT modem and router wifi supplies, all 1.5A 12V jobs. Hope all well with you Andy, I'm distracting myself trying to learn how to use Kicad design software and catching up on the roundtuit list. T'other Andy. |
29th Oct 2020, 4:55 pm | #13 |
Hexode
Join Date: Apr 2007
Location: W Yorks, UK.
Posts: 406
|
Re: SMPSU + Linear regulator combo?
I know exactly what you mean, and I have never been satisfied with that explanation. 'Injecting noise into ground' doesn't mean anything. To have noise, a voltage must be developed between two points, and the question must be answered, how does that voltage develop? Just 'injecting something into ground' is not an explanation. The problem is particularly mysterious because it can still happen even when the SMPS is isolated from mains earth, and the audio system is also not earthed, so there are no ground loops. In theory there would be no path for noisy switching currents to exit the PSU, enter the audio system, and return to the PSU. And yet the noise remains, far louder than obvious things like stray capacitance or simple PSU ripple should allow. I have yet to see the phenomenon explained.
Last edited by daviddeakin; 29th Oct 2020 at 5:01 pm. |
29th Oct 2020, 5:28 pm | #14 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
|
Re: SMPSU + Linear regulator combo?
Yes - it'll work - and is standard practice in many fields.
It's quite common to have a big SMPS that feeds its DC outputs to individual linear-regulators mounted on each of the various circuit-boards in the equipment. This has several advantages: 1] The SMPS regulation doesn't need to be that good because the individual linear regulators on the boards will clean things up. 2] Less likelihood that noise from one board will feed-back down the power-supply rails and perturb the operation of other boards. 3] Heat dissipation is distributed more-evenly throughout the equipmene, making cooling simpler and reducing the number of 'hot spots'. 4] Can make "soft-start" arrangements easier. I'm quite happy using a SMPS essentially as an 'electronic transformer' then using 317s or similar for higher-precision on-board-regulation. |
29th Oct 2020, 6:57 pm | #15 | |
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,086
|
Re: SMPSU + Linear regulator combo?
Quote:
The switching transients in these PSUs put them into the RF band. The typical source of emc are loops in the powerful part of the circuit. So what you get is the earh wire, or whatever, being driven like an end-fed aerial. If you are looKing for your "loop" you have to think beyond wire. RF things are effectively completing their circuit by capacitance - no wire required! |
|
30th Oct 2020, 9:12 am | #16 | |
Hexode
Join Date: Apr 2007
Location: W Yorks, UK.
Posts: 406
|
Re: SMPSU + Linear regulator combo?
Quote:
https://www.youtube.com/watch?v=n_b_0BWP0mI&t=371s |
|
30th Oct 2020, 9:29 am | #17 | ||
Octode
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
|
Re: SMPSU + Linear regulator combo?
Quote:
Essentially the problem here is thinking that something called "ground" actually exists - in the sense that it has zero potential across it. It can be hard to appreciate that a solid lump of metal, for instance an aluminium chassis can be at different potentials on opposite sides, something that most people's understanding of "ground" precludes. But with enough circulating current (for whatever reason) at any frequency some potential difference will exist, even if its only a few microvolts. And in some systems - particularly those demanding very high S/N ratios, that potential can lead to the sort of problem you are referring to. Richard |
||
30th Oct 2020, 9:38 am | #18 | ||
Dekatron
Join Date: Aug 2003
Location: near Reading (and sometimes Torquay)
Posts: 3,086
|
Re: SMPSU + Linear regulator combo?
Quote:
Once you come to accept that "ground" doesn't actually exist, your designs usually improve considerably. Quote:
|
||
30th Oct 2020, 10:46 am | #19 |
Rest in Peace
Join Date: Apr 2008
Location: London, UK.
Posts: 2,508
|
Re: SMPSU + Linear regulator combo?
I routinely work with this configuration, and also with SMPSUs and DC-DC converters in analogue signal applications. As mentioned it is a popular technique but it is not a panacea for all SMPSU hash.
An important consideration especially where the load is floating, is that no amount of linear post-regulation will remove common-mode noise originating from the SMPSU, which is a frequent source of buzz and other maladies. The entire load circuit, with its nice smooth DC, can end up flapping about at hundreds of kHz referenced to something like half mains voltage wrt. true earth, complete with some AM at 100Hz from the switcher duty cycle. If the application circuit were truly symmetrical including all strays, the common-mode noise would not affect its operation; but as soon as there is a ground plane or other major assymmetry (even a linear post-regulator!) it resurfaces as single-ended noise within the circuit. Not unlike modulation hum in a radio that has perfectly smooth HT, due to RF side-effects at the source of that smooth HT. |
30th Oct 2020, 4:30 pm | #20 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: SMPSU + Linear regulator combo?
In the end I knocked the idea on the head, but will file the idea away. A lot of folk over on DIYaudio use SMPSU's in valve amps with success, both for HT and heater supplies, but from reading the above it seems the idea is full of potential pitfalls. I suppose that it's further complicated by all the various SMPSU types and configurations. We're entering the world of voodoo and RF which to some of us is the same thing.
Thanks again for all your IP., Andy.
__________________
Curiosity hasn't killed this cat...so far. |