I'm thinking of using a class d amp for useage in a guitar sustainer (it's what the commercial guitar sustainers use)....for those that aren't familiar, a guitar sustainer is essentially a low bandwidth audio amp, driving an indutive load (an electromagnet) to keep a guitar string vibrating forever!
I'm worried about the phase shift.....I've a hunch that if a class d amp introduces significant (and differing) phase shifts for frequencies upto 1.5khz, then the sustainer won't work well.
so...
1, With known load inductance...how can I calculate the phase shift for frequencies up to 1.5khz for a class d output.
2. What methods (if any) can I employ to keep the phase shift marginal/low upto 1.5khz audio frequency?
3. I'd like to go filterless, so how can I work out the best switching frequency should be for a known load (eg if say the inductive load is 2.5mh...what would be the optimum switching frequency so that the load appears inductive? II'm hoping that because a sustainer driver coil is way more inductive than a typical audio lodspeaker, that this approach is do-able)
thanks,
pesky
I'm worried about the phase shift.....I've a hunch that if a class d amp introduces significant (and differing) phase shifts for frequencies upto 1.5khz, then the sustainer won't work well.
so...
1, With known load inductance...how can I calculate the phase shift for frequencies up to 1.5khz for a class d output.
2. What methods (if any) can I employ to keep the phase shift marginal/low upto 1.5khz audio frequency?
3. I'd like to go filterless, so how can I work out the best switching frequency should be for a known load (eg if say the inductive load is 2.5mh...what would be the optimum switching frequency so that the load appears inductive? II'm hoping that because a sustainer driver coil is way more inductive than a typical audio lodspeaker, that this approach is do-able)
thanks,
pesky
Last edited:
Unless you have some clever digital filters you can't have narrow bandwidth and no phase shift. Filterless Class D can generate RF interference, so may affect any radio mikes you have. An inductor will of course give you a 90 degree phase shift. Why Class D? Do you need particularly high power and efficiency?
I'm not an expert on either Class D or guitars, but I suspect you need to do some more reading.
I'm not an expert on either Class D or guitars, but I suspect you need to do some more reading.
Guitar sustainers are power hogs, and they're typically installed inside a guitar & powered off a 9V battery...therefore any efficiency win back is welcome (hence class D). The fact that commercial sustainers use class D amps, illustrates...
1. That's it's (likely) the way to go (they seldom do this stuff for fun/love)
2. That it can be done.
I guess what I'm after here is someone that knows how to calculate the phase shifts likely to be seen. I'm figuring here that if the switching frequency is so high that a low pass filter can be used that has negligible phase shift at frequencies up to 1.5Khz.
there's a degree of irony in that statement! In the light that you're neither a class d expert, a guitar expert, or a sustainer expert....you're contributing to a thread where all 3 are important & then telling me to do more reading?!!
Last edited:
OK, I understand why you need high efficiency.
Phase shift is not just a matter of inductance but also resistance. What do you mean by 'negligible phase shift'? 1 degree, 10 degrees, anything less than 45 degrees? The biggest source of phase shift will be the guitar string itself, as it is a high Q resonator. Therefore a little phase shift in the circuit will not change the frequency by much.
If I were to build a sustainer I would need to do a lot of reading on guitars, guitar coils etc. I freely admit that. My guess is that you still need to do some more reading on passive filters. Class D amps can provide low phase shift up to 20kHz and beyond, as otherwise they could not be use in serious audio systems. Why do you anticipate that 1.5kHz might present a problem? Is there some reason why you need a particularly low switching frequency?
To calculate the phase shift into an inductor at a particular frequency you just take the tangent of the ratio of the inductive reactance to the driving source resistance. If you already knew that then why ask. If you didn't know that then more reading is needed. You might not believe this, but I am trying to be helpful.
Phase shift is not just a matter of inductance but also resistance. What do you mean by 'negligible phase shift'? 1 degree, 10 degrees, anything less than 45 degrees? The biggest source of phase shift will be the guitar string itself, as it is a high Q resonator. Therefore a little phase shift in the circuit will not change the frequency by much.
If I were to build a sustainer I would need to do a lot of reading on guitars, guitar coils etc. I freely admit that. My guess is that you still need to do some more reading on passive filters. Class D amps can provide low phase shift up to 20kHz and beyond, as otherwise they could not be use in serious audio systems. Why do you anticipate that 1.5kHz might present a problem? Is there some reason why you need a particularly low switching frequency?
To calculate the phase shift into an inductor at a particular frequency you just take the tangent of the ratio of the inductive reactance to the driving source resistance. If you already knew that then why ask. If you didn't know that then more reading is needed. You might not believe this, but I am trying to be helpful.
I think you've misunderstood...I don't want the switching frequency to be 1.5khz....clearly the switching frequency should be very high (100s of khz)....the 1.5khz figure is the maximum fundamental note on a guitar. Again....I need the phase shift for audio signals upto 1.5khz to be minimal through a class D & its output filter..
This is not a question about guitar sustainers (I've built loads that work fine with class AB amps), but more how to apply a class D amplifier to a highly inductive guitar sustainer coil.
Yes, there's already a phase shift between guitar pickup & sustainer driver.....but adding in more phase shift by way of class d induced phase shift IMHO simply adds to the problem...which is why I seek a minimal phase shift (ideally under 1%) for audio frequencies up to 1.5khz.
re reading about passive filters...the waters are muddied somewhat in that there isn't much on the net to read about how a PWM signal output (ie from a class D amp) might be phase shifted vs the input signal.....for most, I'd imagine this info matters not one iota...but for a guitar sustainer it becomes vital.
The whole premise of a user forum falls down with that sentiment! I didn't know which is why I'm here...and forums are where folks come to get 'leads' .....not to be just told to go & read stuff (else why have a forum?!)
This is not a question about guitar sustainers (I've built loads that work fine with class AB amps), but more how to apply a class D amplifier to a highly inductive guitar sustainer coil.
Yes, there's already a phase shift between guitar pickup & sustainer driver.....but adding in more phase shift by way of class d induced phase shift IMHO simply adds to the problem...which is why I seek a minimal phase shift (ideally under 1%) for audio frequencies up to 1.5khz.
re reading about passive filters...the waters are muddied somewhat in that there isn't much on the net to read about how a PWM signal output (ie from a class D amp) might be phase shifted vs the input signal.....for most, I'd imagine this info matters not one iota...but for a guitar sustainer it becomes vital.
The whole premise of a user forum falls down with that sentiment! I didn't know which is why I'm here...and forums are where folks come to get 'leads' .....not to be just told to go & read stuff (else why have a forum?!)
Last edited:
No, I realise that 1.5kHz is not the switching frequency but your highest expected signal frequency. My understanding is that the Class D output is unlikely to be significantly shifted with respect to the input, otherwise stereo would not work too well. PWM might add a little phase shift, depending on how it is done.
Could you clarify what you mean by 1% phase shift? If you want little phase shift then you will need wide bandwidth, yet you originally spoke of low bandwidth.
To calculate the output filter you need to know the effective output impedance of your amplifier. This will depend on the particular Class D method you use. Have you looked at some of the Class D chips used for output from personal stereos, for example?
Hopefully someone will pop up here who has done what you are attempting.
Could you clarify what you mean by 1% phase shift? If you want little phase shift then you will need wide bandwidth, yet you originally spoke of low bandwidth.
To calculate the output filter you need to know the effective output impedance of your amplifier. This will depend on the particular Class D method you use. Have you looked at some of the Class D chips used for output from personal stereos, for example?
Hopefully someone will pop up here who has done what you are attempting.
I mean for all frequencies between 82.4Hz (lowest note on a std tuned electric) to 1.5khz (highhest note (note: I'm talking fundamentals, not harmonics here) I want the phase shift top be less than 1%....'class d input' vs 'effective signal across the load'.
Now with an AB amp this is easy to measure....scope the amp input....scope the amp output ....job done. But becuase a class D output is PWM, I can't go this way.
Sure I could add in a filter & scope after the filter, but I would rather go with a filterless solution (on account the driver is way more inductive than a typical loud speaker....and therefore this should be perfect)
Re the bandwidth of the amp.....I didnt mean the characteristic of the amp had to be low bandwidth....I was referring to the range of frequencies a sustainer amp needs to cater for vs a typical full bandwidth audio amplifier.
Even though the sustainer coil is more inductive than a voice coil I would use at least SOME output filtering in order to avoid EMC problems (you are MAINLY generating a magnetic stray filed after all). If you use a simple pre filter feedback class-d amp then you can equalise quite a bit of the phase-shift away.
With post-filter feedback it is possible to make an amp that has very low phase shift overall to begin with.
Regards
Charles
With post-filter feedback it is possible to make an amp that has very low phase shift overall to begin with.
Regards
Charles
Interesting project. I've nedver heard of such a thing till now, but for discussion's sake...
A class-d amp shouldn't have much phase shift at those frequencies should it? Unless you switch at too low a frequency, but why do that?
Filter less? Not sure that's a good idea with guitar strings being antennaes sending HF muck all around. Even with filter I'd think you need to shield against crosstalk to the pickups.
What is the impedance and inductance of a typical driver? Have you scope'd the signal going into the driver/out of the amp to see how it looks?
There are lots of harmonics in guitar pickups, but I'd thik the driver only really wants to see the fundamental as you have hinted, but filtering out the harmonics will give phase shifts. Then again, is a little phase shift so bad? I'd think it's much worse having all those higher frequencies dancing around and being sent out disturbing your pickups and other stuff. The guitar strings are perfect antennaes.
A class-d amp shouldn't have much phase shift at those frequencies should it? Unless you switch at too low a frequency, but why do that?
Filter less? Not sure that's a good idea with guitar strings being antennaes sending HF muck all around. Even with filter I'd think you need to shield against crosstalk to the pickups.
What is the impedance and inductance of a typical driver? Have you scope'd the signal going into the driver/out of the amp to see how it looks?
There are lots of harmonics in guitar pickups, but I'd thik the driver only really wants to see the fundamental as you have hinted, but filtering out the harmonics will give phase shifts. Then again, is a little phase shift so bad? I'd think it's much worse having all those higher frequencies dancing around and being sent out disturbing your pickups and other stuff. The guitar strings are perfect antennaes.
If you switch directly into a coil of several mH the output will be pretty much pure squares with loads of HF. Mechanically the string's mass will filter out the HF, but electronically you will broadcast to everyone around you. Perhaps upset the soundman's class-d powered active speakersystem...?
I haven't posted for a while, but just have to say that this is a really fascinating application that seems perfect for class-d. Does this work on a string by string basis or is it just driving all the strings?
I don't think you're going to have a problem with phase shift. As mentioned, class-d can take the feedback before or after the output filter, so the filter doesn't add that much phase, not enough to effect the amps own feedback that much. A bad filter design could cause oscillation, but we're not talking extreme high fidelity here.
Your idea of going filterless seems fine. The coil itself should work fine as a filter. I expect the coil is shielded, this would cut down on interference. The strings themselves are going to add to the inductance as the coil is meant to drive them. Again, not hi-fi here, so just getting close I think would be fine.
Given that, why not hook the amp up to your coil, scope the in and out, and check it? Compare it to one of your class-ab designs. It will likely hit it right on the mark. That would be sweet.
Now, just imagine that you added some extra circuitry to control phase at the frequency the string is vibrating at. You could have it also damp the strings. You could go from infinite sustain to a banjo pluck!
I don't think you're going to have a problem with phase shift. As mentioned, class-d can take the feedback before or after the output filter, so the filter doesn't add that much phase, not enough to effect the amps own feedback that much. A bad filter design could cause oscillation, but we're not talking extreme high fidelity here.
Your idea of going filterless seems fine. The coil itself should work fine as a filter. I expect the coil is shielded, this would cut down on interference. The strings themselves are going to add to the inductance as the coil is meant to drive them. Again, not hi-fi here, so just getting close I think would be fine.
Given that, why not hook the amp up to your coil, scope the in and out, and check it? Compare it to one of your class-ab designs. It will likely hit it right on the mark. That would be sweet.
Now, just imagine that you added some extra circuitry to control phase at the frequency the string is vibrating at. You could have it also damp the strings. You could go from infinite sustain to a banjo pluck!
If going filter-less into an inductor, the output of the amp will be a train of squares with a sheit load of HF radiating all over. Go ahead and try, the driver will only respond to the audio frequencies so it'll probably appear to work fine, but the HF RFI noise can potentially ruin someone's day.
What switching frequency are you thinking about? At least pick one where the 25"+- length of antennae you have on the guitar is not 1/4 wavelength.
What switching frequency are you thinking about? At least pick one where the 25"+- length of antennae you have on the guitar is not 1/4 wavelength.
found these on another thread here, running filterless info,
http://www.ti.com/lit/an/sloa023/sloa023.pdf
Class D Amplifiers: Fundamentals of Operation and Recent Developments - Application Note - Maxim
In this app the amp is in the guitar body, and everything is shielded, looks perfect to try.
http://www.ti.com/lit/an/sloa023/sloa023.pdf
Class D Amplifiers: Fundamentals of Operation and Recent Developments - Application Note - Maxim
In this app the amp is in the guitar body, and everything is shielded, looks perfect to try.
I think 1% phase shift is a bit ridiculous to go after - you can tolerate tends of degrees of phase shift in a guitar sustainer. With more phase shift you might need a bit more gain to maintain oscillation, but that's simple to add.
I wouldn't go filterless for one reason - if you hold your sustainer above the magnetic pickup of your guitar, you'll couple the filterless waveform directly into the pickup and send it into your effects chain. Granted there's tone controls/etc in the middle of all of this, but a hundreds-of-KHz signal could make some equipment act a bit squirrely, or it might alias into an audible tone at the ADC of a DSP based effect.
I wouldn't go filterless for one reason - if you hold your sustainer above the magnetic pickup of your guitar, you'll couple the filterless waveform directly into the pickup and send it into your effects chain. Granted there's tone controls/etc in the middle of all of this, but a hundreds-of-KHz signal could make some equipment act a bit squirrely, or it might alias into an audible tone at the ADC of a DSP based effect.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.