Forums

Devil's Staircase

Started by John Larkin May 31, 2019
søndag den 2. juni 2019 kl. 04.05.06 UTC+2 skrev John Larkin:
> On Sat, 1 Jun 2019 18:50:48 -0700 (PDT), tabbypurr@gmail.com wrote: > > >On Saturday, 1 June 2019 17:36:48 UTC+1, John Larkin wrote: > > > >> We're selling a lab-grade voltage source, so we want the output to be > >> stiff. I estimated the 20 mohms primary DCR by scaling from a smaller > >> toroid that we have. The AC output impedance is a separate issue from > >> the DC primary current hazard. > >> > >> We will be sensing output current to control the complex output > >> impedance, so we can in theory tune the box Zout to zero by canceling > >> the transformer impedance. That approaches building an oscillator, but > >> we might take out some of the native impedance. > >> > >> But alternators are very inductive, and our main goal is to simulate > >> alternators. > > > >How do you intend to simulate the voltage spike when they suddenly reduce load? > > > > > >NT > > A bit of our simulated inductance will be a real inductor. It can do > the fast stuff, and our DSP loop can take over behind that. > > It's similar to making a super-wideband inductor by making a series > string of little and big inductors. People do that in bias tees and > such. > > The real-life load on the alternator will probably be shunt regulator, > bridge rectifier, capacitor, so really fast spikes probably wouldn't > matter much. > > We're trying to get them to lend us an alternator. We could spin that > somehow and see how it behaves. >
https://youtu.be/D4DqyElNFAs ;)
On Sunday, 2 June 2019 03:05:06 UTC+1, John Larkin  wrote:
> On Sat, 1 Jun 2019 18:50:48 -0700 (PDT), tabbypurr wrote: > >On Saturday, 1 June 2019 17:36:48 UTC+1, John Larkin wrote:
> >> We're selling a lab-grade voltage source, so we want the output to be > >> stiff. I estimated the 20 mohms primary DCR by scaling from a smaller > >> toroid that we have. The AC output impedance is a separate issue from > >> the DC primary current hazard. > >> > >> We will be sensing output current to control the complex output > >> impedance, so we can in theory tune the box Zout to zero by canceling > >> the transformer impedance. That approaches building an oscillator, but > >> we might take out some of the native impedance. > >> > >> But alternators are very inductive, and our main goal is to simulate > >> alternators. > > > >How do you intend to simulate the voltage spike when they suddenly reduce load? > > > > > >NT > > A bit of our simulated inductance will be a real inductor. It can do > the fast stuff, and our DSP loop can take over behind that. > > It's similar to making a super-wideband inductor by making a series > string of little and big inductors. People do that in bias tees and > such. > > The real-life load on the alternator will probably be shunt regulator, > bridge rectifier, capacitor, so really fast spikes probably wouldn't > matter much. > > We're trying to get them to lend us an alternator. We could spin that > somehow and see how it behaves.
I'd think you'd need that to determine L, C etc. Your app may have that load, but if you make it do under 50Hz it's going to get used for other tasks. Valves are inherently sledgehammerproof, as long as you add a spark gap. Many EEs have failed to achieve that with silicon. Dropping a heavy load on a shaker table could produce interesting 'signals'. NT
On Sat, 01 Jun 2019 08:44:48 -0700, John Larkin
<jjlarkin@highlandtechnology.com> wrote:

>> >>Or how about a high inductance but low resistance external solenoid >>across toroid primary ? Apparently the frequency is in the 400 Hz >>ballpark, so even less solenoid inductance should suffice, compared to >>50/60 Hz. > >I'm guessing that the TI amp would fight to maintain its output offset >voltage. A series cap would be a lot easier. I could quit bitching and >use some big caps. Probably don't need the diodes.
The output LC LFPs are outside of any feedback loops, so the LPF filter inductor DC resistances are in series with the load DC resistances, thus limiting the DC current.
On Sat, 01 Jun 2019 19:16:39 -0700, Lasse Langwadt Christensen wrote:
>> > https://youtu.be/D4DqyElNFAs ;)
There appears to be a cat wandering about under the engine. Surely that's a safety fail? -- This message may be freely reproduced without limit or charge only via the Usenet protocol. Reproduction in whole or part through other protocols, whether for profit or not, is conditional upon a charge of GBP10.00 per reproduction. Publication in this manner via non-Usenet protocols constitutes acceptance of this condition.
On 01/06/2019 00:44, John Larkin wrote:
> On Fri, 31 May 2019 16:26:01 -0700 (PDT), Klaus Kragelund > <klauskvik@hotmail.com> wrote: > >> Why would you add that output decoupling capacitor when it is not needed? > > To maybe avoid saturating the transformer. A good toroid has low DCR. > I could easily circulate a few amps DC in the primary. > >> >> None of the designs I have seen has a capacitor. Dig into it and find that you are chasing ghosts and just making a bad design by add a highly unlinear device >> >> Notice that Class D is all about removing or compensating non linearities of the output stage. >> >> Cheers >> >> Klaus >
If the transformer primary is center-tapped you could watch the tap walking away from midpoint to trap imbalance but it seems a kind of round-about way. Just another idea tossed into the cook-pot! piglet
On 2019-06-01, John Larkin <jjlarkin@highlandtechnology.com> wrote:
> On Sat, 1 Jun 2019 00:42:45 -0700 (PDT), Phil Allison ><pallison49@gmail.com> wrote: > > > We're selling a lab-grade voltage source, so we want the output to be > stiff.
An alternator is a current source, non-ideal due to the gap and other physical constraints, but essentially it is a current source. Jim T. explanind this here several years ago and it stuck with me. After actualy thinking about how an alternator works, I could see that it could only ever be a current source. consider what drive you'd need if you were to replace the rotor with a sationary core and wanted the same output waveforms.
> I estimated the 20 mohms primary DCR by scaling from a smaller > toroid that we have. The AC output impedance is a separate issue from > the DC primary current hazard. > > We will be sensing output current to control the complex output > impedance, so we can in theory tune the box Zout to zero by canceling > the transformer impedance. That approaches building an oscillator, but > we might take out some of the native impedance.
Why not start with a current source driving the primary of the transformer and simulate extra the leakage inductance caused by the gap.
> But alternators are very inductive, and our main goal is to simulate > alternators.
control current in the fast inner loop. control voltage in the slow outer loop. -- When I tried casting out nines I made a hash of it.
On 01/06/2019 10:33, John Larkin wrote:
> On Fri, 31 May 2019 17:15:14 -0700 (PDT), Lasse Langwadt Christensen > <langwadt@fonz.dk> wrote: > >> l&oslash;rdag den 1. juni 2019 kl. 01.55.43 UTC+2 skrev John Larkin: >>> On Fri, 31 May 2019 23:38:58 -0000 (UTC), Cursitor Doom >>> <curd@notformail.com> wrote: >>> >>>> On Fri, 31 May 2019 15:36:39 -0700, John Larkin wrote: >>>> >>>>> I want to make a class-D audio amp, 150 watts or so, using a TI TPA3255 >>>>> maybe. It's good for 600 watts mono! >>>> >>>> Why class D? Doesn't seem like an obvious first choice for audio unless >>>> it's for a PA system or something of that sort. >>> >>> We plan to do three channels, 150 watts each, in a 2U rackmount >>> chassis. >>> >>> It will simulate a 3-phase PM alternator hung on the gearbox of a jet >>> engine. We'll be able to program frequency, voltage, and complex >>> output impedance. >>> >> >> why do you need a transformer? > > For isolation, to make 3-phase and other configurations, and to get > various output voltages. > > Some of the downstream voltage regulators are weird. People like to > short out PM alternators for regulation. Our box will have to tolerate > that. > >> >> and why not some COTS amplifier, you can a 4 channel several 100watt amplifier in 2U for a few hundred $ > > We'll probably do our own class-D modules. We could use one of the TI > eval boards, but they are big and have a zillion jumpers and goofy > connectors. > > We can get a 48V, 600 watt MeanWell power supply for $75! > > We were thinking about putting the TPA3255 on the bottom of the board > and heat sinking it to the bottom of the box, which avoids a fan. It > has a huge, grounded power pad on the top of the chip. > > Isn't most audio power class D now? >
Have you seen these? The THD, IMD and noise specs are quite good: https://www.hypex.nl/oem-audio-amplifiers/
On Sun, 02 Jun 2019 10:32:53 +0300, upsidedown@downunder.com wrote:

>On Sat, 01 Jun 2019 08:44:48 -0700, John Larkin ><jjlarkin@highlandtechnology.com> wrote: > >>> >>>Or how about a high inductance but low resistance external solenoid >>>across toroid primary ? Apparently the frequency is in the 400 Hz >>>ballpark, so even less solenoid inductance should suffice, compared to >>>50/60 Hz. >> >>I'm guessing that the TI amp would fight to maintain its output offset >>voltage. A series cap would be a lot easier. I could quit bitching and >>use some big caps. Probably don't need the diodes. > >The output LC LFPs are outside of any feedback loops, so the LPF >filter inductor DC resistances are in series with the load DC >resistances, thus limiting the DC current.
Right, they add to the output DCR loop. I do want to keep them small, ideally surface mount and low DCR. The TI chip in mono mode wants 4 inductors with pairs essentially in parallel. The 4 inductors might add 10 mohms or so to the loop. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On Sat, 1 Jun 2019 19:16:39 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

>s&#2013266168;ndag den 2. juni 2019 kl. 04.05.06 UTC+2 skrev John Larkin: >> On Sat, 1 Jun 2019 18:50:48 -0700 (PDT), tabbypurr@gmail.com wrote: >> >> >On Saturday, 1 June 2019 17:36:48 UTC+1, John Larkin wrote: >> > >> >> We're selling a lab-grade voltage source, so we want the output to be >> >> stiff. I estimated the 20 mohms primary DCR by scaling from a smaller >> >> toroid that we have. The AC output impedance is a separate issue from >> >> the DC primary current hazard. >> >> >> >> We will be sensing output current to control the complex output >> >> impedance, so we can in theory tune the box Zout to zero by canceling >> >> the transformer impedance. That approaches building an oscillator, but >> >> we might take out some of the native impedance. >> >> >> >> But alternators are very inductive, and our main goal is to simulate >> >> alternators. >> > >> >How do you intend to simulate the voltage spike when they suddenly reduce load? >> > >> > >> >NT >> >> A bit of our simulated inductance will be a real inductor. It can do >> the fast stuff, and our DSP loop can take over behind that. >> >> It's similar to making a super-wideband inductor by making a series >> string of little and big inductors. People do that in bias tees and >> such. >> >> The real-life load on the alternator will probably be shunt regulator, >> bridge rectifier, capacitor, so really fast spikes probably wouldn't >> matter much. >> >> We're trying to get them to lend us an alternator. We could spin that >> somehow and see how it behaves. >> > >https://youtu.be/D4DqyElNFAs ;)
I'm thinking we could just chuck it in one of our milling machines and spin it up. That guy did an awful lot of machining. And talking. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On Sun, 2 Jun 2019 12:51:45 -0000 (UTC), Jasen Betts
<jasen@xnet.co.nz> wrote:

>On 2019-06-01, John Larkin <jjlarkin@highlandtechnology.com> wrote: >> On Sat, 1 Jun 2019 00:42:45 -0700 (PDT), Phil Allison >><pallison49@gmail.com> wrote: >> >> >> We're selling a lab-grade voltage source, so we want the output to be >> stiff. > >An alternator is a current source, non-ideal due to the gap and >other physical constraints, but essentially it is a current source. >Jim T. explanind this here several years ago and it stuck with me. > >After actualy thinking about how an alternator works, I could see that >it could only ever be a current source. consider what drive you'd >need if you were to replace the rotor with a sationary core and wanted >the same output waveforms. > >> I estimated the 20 mohms primary DCR by scaling from a smaller >> toroid that we have. The AC output impedance is a separate issue from >> the DC primary current hazard. >> >> We will be sensing output current to control the complex output >> impedance, so we can in theory tune the box Zout to zero by canceling >> the transformer impedance. That approaches building an oscillator, but >> we might take out some of the native impedance. > >Why not start with a current source driving the primary of the >transformer and simulate extra the leakage inductance caused by the gap.
The class D amps are inherently low impedance out. We's drive each amp from an FPGA and a DAC, to synthesize our sine waves. We'd sense final (after the transformer) voltage and current, digitize, and stuff that back into the FPGA. Then we can synthesize complex output impedance, anything from a voltage source to an inductive current source. Should be fun to Spice. -- John Larkin Highland Technology, Inc lunatic fringe electronics