Dear God - I wonder if you would've changed your thought process if you actually were present at the tests!
The issue is - IMO - why does a 48v SLA PS sound better than the SMPSU-plugged-into-an-isolating-transformer? (Which sounded better than the SMPSU without the isotran.)
The question simply is ... will a steep-slope LP filter (after the SMPSU) cause the SMPSU-plugged-into-an-isolating-transformer to sound as good as - or, hopefully, better - than the 48v SLA supply?
Possibly, it is as simple as what you said ("The source impedance of the supply powering the speed controller") ... except I think you really meant to say "The output impedance of the supply powering the speed controller ".
Regards,
Andy
Last edited:
I use SLAs for my SOTA. Its a 24 volt DC motor. I've got 2 SLA in series, and i don't know why it would cost $500 to implement a similar solution. Twice the batteries, but its a one off investment for listening pleaseure. I also use batteries for my j-fet phono stage. Nice and quiet. I've seen power supplis that are more complex thatn the actual amplifier they power; well this is as simple as. And quiet as. I like batteries.My issue:
* I have an AC motor speed controller on my TT which is powered by a +48v Meanwell SMPS.
* I did an experiment where I used a 48v SLA supply to power the speed controller, instead of the Meanwell.
* The music sounded better as a result!
People have suggested this is because a SMPS has HF hash on the +DC output ... and this has a negative effect on SQ.
Obviously, I could use SLAs on a permanent basis for the AC motor speed controller but this would be quite an expensive solution (~$500, including batteries, case, charger etc.).
Someone suggested a brick wall filter (at, say, 2 or 3KHz) on the DC output of the SMPS would solve the problem - as this would stop the HF hash from getting into the speed controller. This would certainly be much cheaper than $500 - so I was hoping someone could tell me how to implement such a brick wall filter.
Thanks,
Andrew
Hazard - given you're in Melbourne, we really should get together!
For one reason ... I also make a JFET phono stage (the 'Muse') which is battery powered! (And, yes - quiet!)
For another - to discuss SLA TT supplies.
Yours is simpler than mine - you have 2x12v SLAs to power a 24v DC motor.
But how do you control DC motor's speed when the voltage produced by a stacked pair of SLAs can be over 25v when fully charged ... and gradually decreases, as you draw current? I thought a DC motor's speed depended on the voltage fed to it?
Me, I have a motor speed controller driving a 24v AC Premotec - the speed controller needs +48v DC.
When I rigged up a PoC SLA supply for the AC motor speed controller, I used 2 parallel strings of 4x12v SLAs - parallel strings to lower the Zout of the battery supply (given that when you stack batteries, you increase Zout).
So 8x12v/1.3AH Panasonic SLAs is $260 (plus postage) - plus a $50 charger. Plus over $200 for a nice (Italian) case to fit it in. Plus some extras. More than $500!
Regards,
Andy
Hi Andy: An SMPS may spit noise backward into the AC mains as readily as forward into the load.
Your experiments leave open the possibility that your other amplification components (line preamp, power amp) are showing the effects or whatever noise the SMPS is dumping back into the AC mains. If so, focusing solely on the output side of the SMPS may leave you short of where you want to be.
Why not play a digital source, listen with the SMPS and turntable powered on, then do the same with the SMPS and turntable disconnected from the AC mains?
Also, keep in mind that many electrolytic capacitors lose capacitance as the operating frequencies increase, so the noise attenuation performance that your filter actually delivers may be different from the calculations or simulations.
hth
Your experiments leave open the possibility that your other amplification components (line preamp, power amp) are showing the effects or whatever noise the SMPS is dumping back into the AC mains. If so, focusing solely on the output side of the SMPS may leave you short of where you want to be.
Why not play a digital source, listen with the SMPS and turntable powered on, then do the same with the SMPS and turntable disconnected from the AC mains?
Also, keep in mind that many electrolytic capacitors lose capacitance as the operating frequencies increase, so the noise attenuation performance that your filter actually delivers may be different from the calculations or simulations.
hth
Thank you for your suggestions, Jcarr.
Yes, ... I thought I'd solved the 'backwards' problem by using an isolating transformer - but one of the other posters here suggested that these don't do this 100%.
Yes, good test; we will have to do this.
Now you've really ruined my day!
So it may turn out that because of this, the battery supply is the best! (So here is a case where expensive is better! )
So I need to build this pi filter so it can be plugged in between the output of the Meanwell and the input socket on the speed controller - so it can be easily swapped in & out. for listening comparisons.
Regards,
Andy
Yes, ... I thought I'd solved the 'backwards' problem by using an isolating transformer - but one of the other posters here suggested that these don't do this 100%.
Yes, good test; we will have to do this.
Now you've really ruined my day!
So it may turn out that because of this, the battery supply is the best! (So here is a case where expensive is better!
)So I need to build this pi filter so it can be plugged in between the output of the Meanwell and the input socket on the speed controller - so it can be easily swapped in & out. for listening comparisons.
Regards,
Andy
Now you've [B said:really [/B]ruined my day!
Sorry
. Let's just say that awareness of potential pitfalls or weakpoints makes it easier to design around or address them.The following thread may be worth reading through (in particular the posts by Conrad Hoffman and Eva).
http://www.diyaudio.com/forums/power-supplies/106648-paralleling-film-caps-electrolytic-caps.html
hth
andyr - what is your speed controller? Is it a synthesized sinewave variable frequency type with some kind of taco feedback to tune the frequency? If the motor is 24 pole then do you get a dominant ripple current drawn from your power source (eg. something like 24x50=1.2kHz ?)
It is worth considered the power source as having a source impedance which not only comes from the mains supply network, but is also influenced by any closely connected shunt electronic components (such as a power supply for a power amp). Harmonic currents flowing due to the motor converter may partly flow through the amp power supply, and not just simply impress a distortion on the mains voltage waveform (as if you had no power amp power supply connected).
If you know there is a dominant ripple frequency being drawn by the motor, and there is also a switchmode converter fundamental frequency involved, and your speed controller is a poor design with respect to managing that ripple and switching noise, and you don't know of a better speed controller to use, then I'd be looking at the speed controller electronics/design for starters, and contemplating a series harmonic trap for the ripple frequency at the speed controller terminals.
A CRO or soundcard spectrum analyser view of the battery current should show up what is present (in a safe manner).
It is worth considered the power source as having a source impedance which not only comes from the mains supply network, but is also influenced by any closely connected shunt electronic components (such as a power supply for a power amp). Harmonic currents flowing due to the motor converter may partly flow through the amp power supply, and not just simply impress a distortion on the mains voltage waveform (as if you had no power amp power supply connected).
If you know there is a dominant ripple frequency being drawn by the motor, and there is also a switchmode converter fundamental frequency involved, and your speed controller is a poor design with respect to managing that ripple and switching noise, and you don't know of a better speed controller to use, then I'd be looking at the speed controller electronics/design for starters, and contemplating a series harmonic trap for the ripple frequency at the speed controller terminals.
A CRO or soundcard spectrum analyser view of the battery current should show up what is present (in a safe manner).
Last edited:
Hi Tim,
My AC motor speed controller is the "Number 9", developed by Steve Tuckett, here in Melbourne. It is his prototype, so it drives only the 24v Premotec (his 'production' version will also drive the 18v & the 110v Premotecs.)
I'm pretty sure the motor is 24 pole but I have absolutely no idea whether I get "a dominant ripple current drawn from your power source".
The Number 9 is brilliant in that it allows several parameters to be set for the motor:
* The Hz - as this controls speed. Frequency can be set to 0.001Hz - and there is a 45 rpm as well as a 33 1/3rd rpm setting.
* The phase angle between the red & blue wires. So I can set the phase angle to deliver minimum motor vibration.
* The voltage fed to the motor. Once you have minimised motor vibration with the optimum phase angle, the way is open to increase the voltage fed to the motor. Increasing voltage:
- increases motor vibration ... but if this is already minimised, it doesn't increase it noticeably.
- increases torque - this has significant SQ benefits!
I run my motor at 32v - only possibly because the design of my TT means that it is in the open air (not enclosed within a plinth) so does not get hot.
In this case, I assume a battery PS - because it is not connected to mains (and neither is the motor) - would be ideal?
Regards,
Andy
From memory, when we did the tests with the battery supply, the speed controller was drawing about 150ma when the motor was switched on and more than 100ma when the motor was switched off - so the actual 24v Premotec draw was 20-30ma.
Why do you ask, Chris?
Regards,
Andy
Hah - yes, it's a difficult concept to get one's head around, when one is talking about the DC power supply to an AC motor speed controller!
We (there were 4 of us in all the listening tests) heard things like:
1. better definition of the harmonic structure of instruments
2. better decay
3. deeper bass & better bass definition
4. cymbals sounded more like 'real'- ie. live - cymbals.
(Undoubtedly, #2-4 follow on from #1!
IOW, the same kind of stuff that Linnies talk about when they compare the 'sound' of various LP12 power supplies (basic, Valhalla, Lingo 1, 2 & 3, and Radikal).
Regards,
Andy
Just wondering why not use a low noise linear regulator? Just seems odd to choose between a cheap(ish) SMPS or expensive battery setup.
I have a few spare tps7a4700 here and a PCB I designed for it to follow after a lm317 and rectifier + snubber which is what I had in mind when I posted, but looking now it only does 20V out. (Tps7a4701 would work but I dont have a PCB for that)
Another option is opc's the wire PSU v2 PCBs (see vendors bazaar thread for noise measurements).
These newer linear reg ICs are better in terms of noise and psrr than the lt108x or lm3x7 types ...
Yet another option would be to use something like a hybrid similar to what's on the positive half of Ti PMP8372 reference design
Chris
I did build a low noise linear regulated PSU, Chris - as I was pretty sure it was going to sound better than the MeanWell SMPSU which Steve Tuckett supplies with his AC motor speed controller.
This was designed by Hugh Dean - and it did sound better ... interestingly enough, though, when we tried 2 further tests:
1. linear regulated PSU with isolating transformer in front of it, and
2. MeanWell SMPSU with isolating transformer in front of it
... suddenly, #2 came out on top!
It was only when we subsequently compared this against a 48v SLA supply that we realised the Meanwell/isotran combo still needed some tweaking.
Hence my search to find a LP filter to chop off the HF hash which a SMPSU produces.As a result of the helpful responses received here, I can now build a suitable pi LC filter - in a form that makes it easy to put it into circuit and then remove it - so we can see whether or not:
* it does improve the sound from the MeanWell/isotran combo, and
* it's better than the 48v SLA supply.
However, taking Jonathan Carr's comments into consideration, it may well be that the pi LC filter doesn't work at the frequencies put out by the SMPSU? (Which would mean that the SLA supply is 'king'!
)Regards,
Andy
Last edited:
Because of this, Jonathan, would it be advantageous to put film bypass caps on the 220uF & 330uF electrolytics in the 5-element pi filter?
AIUI, these smaller values allow there to be some capacitance at high frequencies - however, their use seems to be a matter of contention? (so don't know whether to do it or not!
).Regards,
Andy
Andy: A few more things that you could try.
1. Put an EMI_EMC filter between the AC mains and whichever power supply you are using (SMPS or non-SMPS). Together with the isolation transformer, this should reduce whatever noise is being injected into the local AC mains. Ideally you'd design or build your own EMI_EMC filter, but ready-made filters are better than nothing.
2. In addition to designing a passive pi filter for the output of the SMPS, you could design a similar filter for the linear PSU - integrated into the rectification-smoothening section and before the linear regulator. Rectifying AC into DC generates wide-band noise, and you shouldn't assume that your regulator circuit will be able to block it all, unless it was explicitly designed to do so, or you assist the regulator with a suitable passive or active filter.
When rectifying 50_60Hz AC mains, the problem-causing noise will start at a lower frequency (obviously) but will only extend up to 50~100kHz. This puts it within range of electrolytic capacitors (at least organic-semiconductor or conductive polymer types). In comparison, the noise from an SMPS will reach well into the FM band, making it much harder to silence.
3. Put an EMI_EMC filter between the AC mains and your line preamp to reduce the amount of noise that can get into the preamp. And if you can, replace your preamp's power transformer with a E-I frame transformer that has split bobbins. The split bobbins will reduce capacitive coupling between primary and secondary windings, again reducing the amount of pollution that reaches the preamp circuits.
1. Put an EMI_EMC filter between the AC mains and whichever power supply you are using (SMPS or non-SMPS). Together with the isolation transformer, this should reduce whatever noise is being injected into the local AC mains. Ideally you'd design or build your own EMI_EMC filter, but ready-made filters are better than nothing.
2. In addition to designing a passive pi filter for the output of the SMPS, you could design a similar filter for the linear PSU - integrated into the rectification-smoothening section and before the linear regulator. Rectifying AC into DC generates wide-band noise, and you shouldn't assume that your regulator circuit will be able to block it all, unless it was explicitly designed to do so, or you assist the regulator with a suitable passive or active filter.
When rectifying 50_60Hz AC mains, the problem-causing noise will start at a lower frequency (obviously) but will only extend up to 50~100kHz. This puts it within range of electrolytic capacitors (at least organic-semiconductor or conductive polymer types). In comparison, the noise from an SMPS will reach well into the FM band, making it much harder to silence.
3. Put an EMI_EMC filter between the AC mains and your line preamp to reduce the amount of noise that can get into the preamp. And if you can, replace your preamp's power transformer with a E-I frame transformer that has split bobbins. The split bobbins will reduce capacitive coupling between primary and secondary windings, again reducing the amount of pollution that reaches the preamp circuits.
Hi Andy: I cannot say for sure, as my use of pi filters has been predominately for 50_60Hz power supplies rather than SMPS. However, I suggest at least making the SMPS pi filter from conductive polymer electrolytics - Nichicon, for example, manufactures these in voltages as high as 125V. If the per-piece capacitance value is too small for the pi filter, I would use multiple pieces of the same value in parallel, possibly with provisions for sub-0.1 resistors in series with each capacitor to dampen resonances.
As bypasses, I would add smd X7R or Y5V ceramic capacitors. Smaller size and better HF characteristics than films. The "gotcha" with ceramic capacitors (excluding COG and NPO types) is that their capacitance may fall with increasing voltage, so there may be a need to derate.
One more thing to watch out for is the parasitic capacitance of the inductors (although you may not find a specification for this). Being wound, an inductor coil will possess a certain amount of leakage capacitance from one end of the winding to the other, and this capacitance will present a short-circuit for high-frequency noise. Some CM chokes are visibly wound in multiple sections, which reduces the parasitic capacitances by placing them in series. Hopefully you can find a plain inductor which is similarly wound in multiple sections. Barring that, you could look for low DCR in comparison to the inductance (low DCR implies fewer windings).
hth
As bypasses, I would add smd X7R or Y5V ceramic capacitors. Smaller size and better HF characteristics than films. The "gotcha" with ceramic capacitors (excluding COG and NPO types) is that their capacitance may fall with increasing voltage, so there may be a need to derate.
One more thing to watch out for is the parasitic capacitance of the inductors (although you may not find a specification for this). Being wound, an inductor coil will possess a certain amount of leakage capacitance from one end of the winding to the other, and this capacitance will present a short-circuit for high-frequency noise. Some CM chokes are visibly wound in multiple sections, which reduces the parasitic capacitances by placing them in series. Hopefully you can find a plain inductor which is similarly wound in multiple sections. Barring that, you could look for low DCR in comparison to the inductance (low DCR implies fewer windings).
hth
Hi Johnathan, hi Andy .
Looking at how the motor works suggests it would do a fine job of being a filter. As others have said it must be for another reason. I have to say if mine a linear supply seems the answer. I suspect it could be very small . 2 x 24 V might be cheaper. How does the $500 come about? I assume it is only the 48 V DC that you need? 4 motorcycle batteries might suit.
If you or a friend could spectrum analyze the supply that would help. Be careful it is inside the permitted voltage range, the 10:1 probe would help.
I have two train controllers I bought in Berlin when visiting a friend. They were very cheap ( 10 Euro ) as they are linear and people wanted SMPS. I thought they might do for non critical things. I was wrong as they work fine for most things. They are based on LM317 which is not at all bad. It proves humble linear works very well.
Could it be the arm and PSU share a ground connection? If so that might be the way in for the RF noise. Or is the PSU near the arm or cartridge?
Looking at how the motor works suggests it would do a fine job of being a filter. As others have said it must be for another reason. I have to say if mine a linear supply seems the answer. I suspect it could be very small . 2 x 24 V might be cheaper. How does the $500 come about? I assume it is only the 48 V DC that you need? 4 motorcycle batteries might suit.
If you or a friend could spectrum analyze the supply that would help. Be careful it is inside the permitted voltage range, the 10:1 probe would help.
I have two train controllers I bought in Berlin when visiting a friend. They were very cheap ( 10 Euro ) as they are linear and people wanted SMPS. I thought they might do for non critical things. I was wrong as they work fine for most things. They are based on LM317 which is not at all bad. It proves humble linear works very well.
Could it be the arm and PSU share a ground connection? If so that might be the way in for the RF noise. Or is the PSU near the arm or cartridge?
Well, I did build a linear PS, Nigel; as I reported earlier, the listening results in order of increasing SQ, so far, are:
* MeanWell SMPSU
* linear PSU
* linear PSU with isotran in front of it
* MeanWell with isotran in front of it
* 48v SLA supply.
Given the battery supply sounds the best, I was searching for a way to 'kill' the HF hash which I know SMPSUs deliver - to see whether the MeanWell would come to the fore. However, it may well be that it is too hard to remove 100% of the HF hash ... so the battery supply will still be king!
$500 comes about as follows:
* 4 x 12v SLAs, stacked, are required to deliver 48v.
* with stacked batteries, you get a Zout for the stack which is 4 x the Zout of an individual battery.
* hence I want to use 2 parallel strings of 4 x 12v SLAs = 8 batteries, to lower the Zout (16 batteries would be the ultimate ... but this a) would be even more expensive and b) would require a much bigger case.
).* batteries are $32 ea, so that's over $300, including a charger.
* a nice case from Mudushop will be over $200 (including some CNC work).
That would be very nice to do; unfortunately I don't know anyone who has such a device.
Humble linear certainly does work very well - but I don't like LM317s and the like. The linear supply which I use was designed for me by an amplifier-designing mate of mine; I also use it to power my active XO and a phono stage. This uses an emitter-follower configuration for regulation - and I know it sounds better than one built with normal regulator chips because I first used it in the active XO, where it replaced a conventional regulator chip PS (Rod Elliott's P05).
The SMPSU is not close to the cartridge - it's on the floor (whereas the TT is on a shelf) ... so it's 3' away.
However, the arm & the SMPSU ultimately may well share a mains earth connection; the earth lead on the arm goes back to the earth terminal on the phono stage - which is connected to the case ... which is connected to mains earth via the earth tag on the IEC socket. The SMPSU has an IEC socket so mains earth is supplied to it ... but whether what's inside the box is actually connected to mains earth ... I have no idea.
Regards,
Andy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.