ST head modification for 48 volt battery charging

[bschwartz]

I'm not sure how much it matters here, but I still have quite noticeable flicker even after I changed from the ST head to the PMG.

[BruceM]

The 5.5 Hz AC voltage variation of a genset driven by a 650 Hz Listeroid is quite pronounced.  With a low pass passive filter to filter out most of the 60Hz, it's easy to see with an oscilloscope.  (see first photo below).

Things that help:  Geno showed that the SOM flywheels make a significant difference in measured flicker, as well as subjectively.  My AVR circuit which has a fast response will reduce it as much or more than SOM flywheels, especially on smaller heads. Both would be even better.

Cliff did some experimenting with my basic AVR design on his ST-5.  He reduced the AC voltage to DC level integrating capacitor to find the optimum flicker reduction (by his eyes).  He got down as low as 2uF, and while voltage regulation suffered, visual flicker continued to improve.

I later tried this on my ST-3 AVR, and found that it made a big difference for me, visually.  I have flicker sensitivity due to brain rot (MS) and epilepsy.  I could not use the Lister for lighting at all with the stock ST-3.  With the AVR and 100uF integrating cap, flicker was much improved, but I could not work in my shop with this lighting, it was too "disturbing".  With my AVR and a  reduced integration capacitor (10uF), I am able to work in my shop with 250 watt incandescent heat lamps, without any problems for long periods and was able to work in the house attic with 100 watt bulbs.  (The latter I wouldn't want to spend more than an hour around.)  Incandescent bulbs of higher wattage have less flicker due to thermal latency.  60watt bulbs were problematic.

The voltage variation due to 5.5 Hz power stroke looks like a rectified, rounded sine wave, with a period of 5.5 Hz.  The low point is the TDC for the compression stroke. (see first photo below) The AVR has a problem correcting for the low point;  there is so much inductance in the rotor that rapid changes in output voltage can't be made.  The AVR tries to compensate but isn't smart enough to anticipate the compression/ignition stroke problem in advance.

The first photo below is the AC of my ST-3 on harmonic with a low pass filter appled.  You can see the 5.5 Hz variation of voltage clearly.  The second is with my AVR, and 100uF for an AC voltage integration capacitor.  You can see the reduction in 10Hz voltage variation is significant, not the change in scale.  I have not measured this with the present 10uF cap, but it is visually as dramatic a changeas was the reduction from stock harmonic to AVR with 100uF integrator.

I think flicker can be reduced even further with "software syncing" of the engine's combustion event, and by the use of even higher voltage excitation, and possibly rotor shunting, with slow average current limiting to protect the rotor windings. Higher voltage excitation will improve response time for dropping AC voltage. Briefly shorting the rotor would improve response time for rising AC voltage.  By "software syncing" I mean adding a microcontroller to track combustion TDC, perhaps by Ronmar's (ingenious) method of electret mic on the high pressure injector line to detect the injection pulse.

We can then calculate when to start applying compensation for the next compression stroke, subtracting this compensation value (analog output) from the integrated AC voltage (now a DC level). This will result in the AVR compensating sooner, anticipating the compression stroke.  Likewise, the ignition stroke can be anticipated precisely, and shunting or shorting of the circulating current in the rotor could be used (briefly) to provide a more rapid reduction in AC voltage.  (No AVRs I've seen do this at present... excitation is just stopped and the inductive circulation of excitation is allowed to dwindle on it's own.)

The wild card is- will it good enough, particularly for low wattage bulbs?  We can correct the AC voltage variation, but we can not change the frequency variation;   while we can achieve 60Hz on average,  the AC frequency will slow during the compression stroke, and get faster after ignition.  
It may be that there is diminishing return- that at some point this frequency variation (which can't be corrected) is noticeable to some people.  5.5 Hz is right in the biological frequency (brain wave) range, and this is a sensitive area for stimulus.

Based on my experience with experimenting on my ST-3, and that I am an epileptic with flicker sensitivity (60 Hz refresh LCD displays bother me as badly as a 60 Hz CRT, even though the depth of modulation of the LCD is less than 1%), I think that flicker compensation probably could be achieved which would be tolerable for me, and thus most "flicker intolerant"  people, at least for the ST-3 and perhaps the ST-5.  The larger heads have too much inductance caused latency.

If we could go even further and affect the head design,  we would reduce the inductance of the rotor, use fewer turns with heavier gauge wire, so that greater excitation energy would be required, but reducing the inductance and latency.

Veggie- what do you think about moving this to a separate thread "ST Heads and Lister Flicker"?
However you want to handle it is fine by me.

[Apogee]

Ok,

Get ready for a left turn...  

I think there is more than one way to skin this cat.

Since we are modifying the head to 60 volts or so, what about just raising the speed a little if needed (not sure we'll need to), rectifying the output, filtering with a bank of caps, and using the filtered output to feed a very beefy mosfet power amplifier circuit.

Then we feed the power amp with a perfect 60 cycle reference source.

The supply voltage could fly all over the place, and as long as the lowest dip was above the needed rail voltage, it wouldn't matter.  No transformer needed because we are already there with the 60 volt supply.

We could even water cool it using the engine cooling loop, thereby recovering any loss as heat from the electrical side!

The amp would be far faster than trying to control the collapse of the magnetic field in the genhead, and the resulting waveform would be damn near perfect.  For that matter, we could include a servo circuit that could track the supply voltage and have it compensate for irregularity as needed.

Perfect AC for inside the house, while at the same time, a concurrent 60 volt supply for charging batteries if needed.  (no need to run the battery charging circuit through the amp, as the raw genhead output would likely be fine for charging)

I wish I could take credit for this basic idea, but I can't.  The medical industry has been regenerating power for a long time.  Also, PS Audio builds their Power Plants for HiFi that use the same basic idea.  PS Audio took it a step further by playing with the input waveforms because they figured out that it affected the sound of the components plugged into it.  By using a microprocessor, they are able to literally divide a waveform in half to optimize charging of power supply circuits.  The amplifier circuit is fast enough to track the constantly changing waveform, and supply it to the output.  Their latest versions are class d amps I believe, and they're getting high output from them with very little heat at about 85% efficiency.  I believe we could do better than that because we wouldn't have the transformer loss and could recover the heat.

We wouldn't need to get that fancy; a simple, clean 1 volt, 60 hz reference source would be all we would need and that would be pretty easy to generate.

Here's what PS Audio is doing for reference:

http://www.psaudio.com/ps/products/description/power-plant-premier_copy?cat=power

This is about the multiwave technology that they came up with (no I don't work for them, just interesting technology I think!):

http://www.psaudio.com/ps/products/description/multiwave-ii-upgrade-kit?cat=cables-accessories

Told you it was a left turn!  I figured you guys needed something to think about while eating your breakfast...

Steve  

PS - And no, this doesn't subscribe to the KISS principle, but it would make for a VERY usable Lister power plant.

[mobile_bob]

ok, how about a right turn right about here...

as  far as i remember the amplifier you refer to can only be about 50% efficient, this theory of operation
has been around for a very long time, and was one of the ways they made pure sine wave several decades ago.

no perhaps an arguement could be made that mosfets would be more efficient that power transistors in the final
drive of the amplifier, but would it improve the efficiency to even 60% or perhaps 70%

maybe, it could be done,
and would be interesting to see what others have to say on the subject, that know a heck of a lot more than me.

bob g

[Apogee]

Well, PS Audio claims they are at 85% and that includes the power supply.

We wouldn't need the transformer so we should be able to do a little better.

Paul McGowan is known as a no BS guy in the audio world.  I can't imagine him claiming that efficiency if he couldn't back it up.  It's in the first link of my previous post.

Since part of our goal is using the excess heat anyway, if the overall system (engine and generating apparatus) dumped a little more heat in trade for high quality power, I didn't see that as a bad thing.  As long as the heat from the amplifier assembly was being used, whether in the cooling loop or in some other way, that part of the system would be close to 100% efficiency.  Obviously if the heat wasn't being used, then we take a X% hit.

I just figured that I'd toss it out there for you guys to think about.

I found Bruce's post very intriguing because detecting the combustion event knock would solve the problem of the constantly changing timing on the Lister generator.  We could also do it using a crank position sensor like I posted.  The advantage of an adjustable sensor would be that we would be able to easily trigger X degrees BTDC which might give enough time to initiate the field change.   If we could figure out a way to change the exciter field fast enough, problem could be solved with his AVR.  That would be, by far, the most elegant solution.

I didn't realize the Changfa ran so fast.  Because of the one to one pulley setup, it would be the perfect test case for combustion event/pole alignment testing. However, the flicker will be masked by its higher speed.  No reason why it couldn't be slowed down to exacerbate the problem for testing though.  Voltage or freq being "incorrect" wouldn't matter for testing, the combustion event issue would still be the same.

Steve

[BruceM]

Steve, the "high end Audio" field is so full of baloney, few real EEs would want to be associated with it.  Being known as a "no BS" guy in that field would be sorta like being a "upstanding" pedophile.

The method you describe would certainly not be 85% efficient.  Linear amplifiers and that kind of efficiency don't mix- it would require a class D (bang-bang) amplifier, which is hardly high end audio.  You may have misunderstood the efficiency claim, or the guy belongs in the latter category above.

But a similar approach to what you are suggesting is viable and is just what some guys like Geno have done- use the Lister to generate bulk DC, which a charge controller can then PWM  into a battery bank and then run a sine wave inverter to generate flicker free AC.  The battery bank provides power for small loads, which the Lister can't do efficiently, as well as power for (short run time) loads larger than the Lister can do, and of course PV can also be used as the charging source.

[Apogee]

Sigh...

Not trying to waste your guy's time...

Seems like you guys already know all about it without even bothering to look at what they're doing.  I guess I shouldn't have posted that it came from the audio world.  I agree that much of the "high end" world is BS.  However, there are also some folks in that industry who practice real engineering, can actually hear, and keep upping the anti.  

I know this works because I bought one of the early models off the used market as an experiment.  It was fascinating how it took the brittleness out of the sound.  Also interesting how the other waveforms affect the sound.  You can easily hear it - it's that blatant.  The change is likely just caused from eliminating all the BS line noise (something you Bruce have been repeatedly talking about), but regardless, it works.  Changed nothing else except plugged equipment into it...

I remember reading tech notes as they were developing it because they had problems keeping the line noise out of the output.  You can't just build a normal amp, feed it an input signal and have clean output if the power supply is dirty.  They chose to engineer in an isolation transformer to keep the noise on one side of the power supply (if that makes sense).  Yeah, they could have used filters, but they tried that first, and found full isolation worked better.  Been ten years since I read this stuff, so my brain is fuzzy on the details...

The second link I posted gets into the waveform technology, and yes, it's different and patented.  I posted the links that I did because I thought you guys would find what they're doing to the waveforms interesting.  Bruce, as an EE you will understand that it took a bit of research to achieve what they're doing.

I did post that I thought it was a class d amp.  Also, this is being used on the power side, not the music side, so not sure using class d in this application is a bad thing.

I have been planning to take Geno's approach for a long time, and have posted such previously.

This came to me as a way (once designed and built) to solve the dirty power/flicker issue for those that didn't wish to run a battery bank/inverter setup.  Yeah, it costs a bit of efficiency, but I felt like the "hit" was almost fully recoverable into the cooling loop if it was done right.

Steve

[Crumpite]

Sigh...

Not trying to waste your guy's time...

I agree that much of the "high end" world is BS.  However, there are also some folks in that industry who practice real engineering, can actually hear, and keep upping the anti.  

I know this works because I bought one of the early models off the used market as an experiment.  It was fascinating how it took the brittleness out of the sound.  Also interesting how the other waveforms affect the sound.  You can easily hear it - it's that blatant.  The change is likely just caused from eliminating all the BS line noise (something you Bruce have been repeatedly talking about), but regardless, it works.  Changed nothing else except plugged equipment into it...

I did post that I thought it was a class d amp.

Also, this is being used on the power side, not the music side, so not sure using class d in this application is a bad thing.

Seems like you guys already know all about it without even bothering to look at what they're doing.  I guess I shouldn't have posted that it came from the audio world.  The second link gets into the technology, and yes, it's different.

I posted the links that I did because I thought you guys would find what they're doing to the waveforms interesting.  Bruce, as an EE you will understand that it took a bit of research to achieve what they're doing.

I have been planning to take Geno's approach for a long time, and have posted such previously (see my post about trying a forklift motor as a 48vdc generator).

This came to me as a fairly easy way (once designed and built) to solve the dirty power/flicker issue for those that didn't wish to run a battery bank/inverter setup.  Yeah, it costs a bit of efficiency, but I felt like the "hit" was almost fully recoverable into the cooling loop if it was done right.

Steve

Steve,

I won't argue that a person can hear differences in different audio gear and with different power supplies.
But without numbers, you are arguing audio 'taste' rather than engineering.
People say, "I think this is true" and others say just the opposite, but now you're talking opinion and religion - it's futile !!!

I did read the links you posted, and noticed a very, very large baloney factor in both links.

There is little/nothing in their designs that isn't already being done everyday in existing power conditioning circuity.
All they are doing it rectifying the AC coming in and then using the DC to reconstruct the AC waveform with very high fidelity.
You'd have to look at their circuit topology to see if it could even handle the power dip in the Lister type AC.
My self, I'd guess not. The seem to be trying to produce a clean AC waveform, not handle an 11 Hz brownout condition.
You *could* design a circuit to do it, but until someone tries their unit under Lister power, you certainly can't bet on it.

You get the same thing by rectifying your AC, feeding it into a battery bank and then feeding an inverter - just like we are already doing...
And you can already get true sine wave inverters.

Engineers always assume that High End Audio is guilty until proven innocent, and for very good reasons !

Let's not argue high end audio, let's argue engineering, Please ?
Daryl

[BruceM]

OK, Steve you caught me, I didn't read the audio company's links.  I didn't want to as I was afraid of what I would find.  

I just did read them, which makes me feel dirty and want to take a shower. I'm covered in marketing bullshit!  You've been had.  

If you had such horribly dirty AC that it was affecting your audio amps sound quality, you should apply basic EMC techniques to find and fix the noise source. It's likely some gear in your own house, if it's that bad.  

If with reasonable AC quality, your audio amplifier has such a poor DC power supply, or has such poor  power supply noise rejection that you are getting audio distortion, then you must have a very poory designed audio amp.  It is at the DC supply where filtering and or regulation should be done.  Adding an inverter before the DC supply to make up for a bad DC supply design is NUTS.

I'd further suggest that these products are likely introducing significant high frequency noise.  If you will put an AM radio tuned between stations near the AC output of the device, you will likely hear it screaming.  

The second link, the "Multiwave II" should be entered for nomination in the "Audiophile Marketing Bullshit" hall of fame.  Adding AC harmonic distortion, by adding an inverter providing a multi-frequency AC waveform is going to make the audio sound, after a DC power supply, better.  Wow!

Is it April 1st already?

[mobile_bob]

audio amplifier aside, it seems to me that one would be better served to just design and build a real inverter
rather than adapt an audio amplifer?

even then it is hard to imagine building an inverter that is of sufficient power, clean waveform, reliable, etc
for the cost of a commercially available product, but maybe?

how am i to rain on anyones parade?

there are also tons of ups systems out there that have exceptionally good waveform, low distortion and are reliable
even though efficiency is in the mid 80's for most of them, they are so inexpensive used it is hard to dismiss them
out of hand and not use one such as Geno has.

they are really very solid performers that are made to work for a very long time.

bob g

[BruceM]

Sorry, I really should be more tactful.  I just loath marketing bullshit and the fleecing of the public by unscrupulous companies.  I shouldn't get so pissed when it's high end audio- the customers can afford to get fleeced, at least, but it's still fraud.

Steve's concept is fine, but as Bob suggests, it makes more sense to use an inverter on DC from the ST head to create flicker free AC, since you need at least 5KW of surge capability. That might fit someone's backup power plans.

Most off grid setups will have batteries and inverters already, and Geno is using them to give him some flicker free power, while charging at up to near full power from his Lister.

[Apogee]

Ok,

Let's leave all of the marketing hype behind.

My point was simply that if we could build or find something that would accept input from the 60 volt genhead and output clean power, it would solve the flicker problem.

Yes, it would involve a hit from an efficiency standpoint, but it might be worth it if it makes the genset more usable.

How we get there, is open.

I like the idea of finding inverters that are already purpose built, but have never seen one that would fit this bill.  That having been said, not sure where to look either.

I was just trying to come up with a solution that could be used without batteries that would respond / be manipulated more quickly than the exciter field in the genhead.

Steve

[BruceM]

The direct to inverter approach would need a 5 or 6KW (surge) sine inverter with an input DC voltage that matches the ST head well, so that a transformer can be avoided. This means 60V or 120V. (With inductive filtering, DC volts =AC volts.)  The 60V is suitable for inverters designed for 48V battery banks, and there are some large UPS inverters which use higher battery voltages. 

I think most problems are best solved at the source, so I wish we had someone interested in pursuing the special Lister Flicker AVR. 

[Jedon]

I have a 120V DC UPS 14KVA, could this be useful? It has a huge transformer and very large capacitors.
http://picasaweb.google.com/jedon13/20080918UPS#

[Crumpite]

I have a 120V DC UPS 14KVA, could this be useful? It has a huge transformer and very large capacitors.
http://picasaweb.google.com/jedon13/20080918UPS#

Holy cow !

I wonder what it's efficiency is ?

Daryl