Of course, you don't use peak current mode control in a half bridge converter. A current source into a coupling capacitor is just as dumb as a voltage source (PWM imbalance) into a PP or bridge inverter. :-)
Peak current mode works nicely for PP (e.g., UCC3808). Curiously, I once breadboarded the equivalent circuit of the '3808 using discrete gates; it suffered from subharmonic instability. Not fatally, it didn't walk -- it can't walk, because PP is DC coupled -- but it did, shall we say, veer to the side of the walkway. Unsure if the '3808 just has better matched timing (likely!), or if it has additional logic to control imbalance.
Skip the transformer, get a proper higher-voltage class D. I like the IRS2092 chip, +/- 100V range. Lots of assembled PCBs on eBay, free shipping from China, two-three weeks. Just search on IRS2092. Choose from among the smallest wimpiest ones.
Assume a sinewave voltage source driving a saturable inductor. The inductor current waveform will be a sine wave, steady-state. If you add a small DC offset to the source, there will be a slow ramp of current superimposed on the current sinewave. This will eventually push the inductor towards saturation. As it approaches saturation, the sine wave gets lopsided, with current peaks in the Isat direction. The combined current waveform can look like a stairstep.
In my case, series capacitance breaks up any long-term current integration. Adding diodes across the cap is probably OK, as it disconnects the amplifier offset and adds volt-seconds in the proper direction to discharge the DC current ramp. If the user applies an out-of-spec low frequency signal, the diodes protect the caps from overvoltage. That's a whole nother issue.
Maybe I could use a supercap, keep the phase shift down. I've been looking for a reason to use one of them.
I'd simulate it, but I don't have a suitable saturable inductor model, and it would take a bunch of time. I'll lay out the board with some options, in case I do get into trouble. It would probably be safe without the diodes.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I measured it with a 50-ohm function generator, by finding the 3dB point. That let me measure it in the low audio range, where it will work. Most L meters will report nonsense when connected to a big power transformer. I measured the primary and secondary independently, and the ratio was very close to n-squared. This is a custom toroid designed to work in the 300-5K Hz range, not a 60 Hz part. Of course L will vary with excitation, but I measured it in about the expected operating range.
404 error.
That's only 500 uF, lots of phase shift. Big caps, too. I want at least a few millifarads. Maybe a supercap is the best thing to use, say 0.47 farad at 5 volts.
And massive phase shift. The max DC offset of the amp is 60 mV and the tranny primary is 70 mohms; that's almost an amp of static current. Filtering the amp input doesn't fix that.
The TI amp claims to be pop and click free. What's the difference between pop and click?
They surge, but the AC line doesn't much care. There's an entire power grid available to demagnetize it.
I want all the power I can get in the volume (and budget) available. And any iron-core transformer will saturate if you put enough DC across it, and 60 mV is "enough" here. It's not a matter of being within 2:1 of safety; it's a couple of orders of magnitude.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
See "Inductor_with_Saturation.zip " on the Device Models & Subcircuits Page of my website.
Doesn't have hysteresis modeled... but as I posted recently, I've devised a way to do Analog hysteresis... adding to the Inductor Model RSN ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I'm looking for work... see my website.
Thinking outside the box...producing elegant & economic solutions.
I tried it on the TI eval board, with 4000 uF plus two diodes in series with the transformer primary. Seems to work fine. I did a bunch of sloppy stuff and didn't damage the TI chip. It does make funny noises some time.
It's putting 33 volts RMS 400 Hz into 25 ohms, 87% efficient after the transformer. Nice chip.
formatting link
I put some questions about this chip on the TI forum. No answers so far.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
I do want isolated output, and if I ever want to do, say, a 120 volt version, I can buy a new transformer.
The transformer lets me use full-bridge drive. Half-bridge can do interesting things driving reactive loads. Full-bridge also lets me run the box off a single 24 volt wart.
I would never use potentially counterfeit Chinese parts in aerospace gear, or actually in anything that we make.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I expect to have to drive reactive loads. Linear amps get very inefficient working in reactive quadrants. I also want to get as much power as I can, over a wide output voltage range, out of a 24 volt wall-wart and a small box without gigantic heat sinks.
A speaker is mostly resistive, so is a good load for a linear amp.
The only problem I'm seeing is the potential for DC error into the transformer primary causing flux buildup, and the series caps fix that. A linear amp would have the same issue.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I already have a Peavey 800 watt stereo amp and some rackmount transformers, but that isn't portable or programmable. And it will eventually die.
We'll probably also offer this as a product, an accessory to our synchro and LVDT processors. I haven't found anything like this on the market; if I did, I'd just buy it.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I'd have to sift milliamps of DC out of amps of varying AC waveforms, and apply a DC correction around the class-D amp somehow. Some series capacitance does that for me.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
As with any good design practice, if one defines all the measurable specs u p front for performance, cost, reliability, environment and time contraints , one can easily define Design Validation Test plans or DVT to rate the des ign result. I used to do this for Magnetic HDD design and used non-destruc tive "margin measurements" to functional failure on critical parameters an d acceptance criteria for pass/fail tests. Test format included setup and diagram , 1 page per test. Summary of all results on 1 page.
You may want to consider this "motus operandi" in your design projects, suc h as this. I learned this way from all my past experiences in Aerospace in 70's and working with Japanese Eng.'s in 80's while at Burroughs/Memorex/U nisys.
Saturation margin depends on many variables , each contributor may be measu red , such as dynamic loads, supply, f, temp, then choose optimal nominal e nergy density, with desired margin. The energy is stored in the distribut ed dielectric gaps between magnetic particles, while the saturation heating occurs in the conductive particles. The Remenance depends on initial cond itions and current*time integration over each cycle alone with current used for feedback can be applied like a PID feedback loop if one wants to get f ancy on increasing energy density stored with feedback for protection, but at what cost?
We don't know the needs of all the people who may use or buy our products. And we don't know how any particular specs will affect sales. We don't know the price elasticity of our products. Some fraction of the things that we design (I thought 50%, others claim
75%) never pay off. So we don't have hard performance metrics. We make a thing as good as we reasonably can without getting crazy, and try to incorporate hooks for improving features or specs if people ask for them. The specs and features evolve as the design progresses, which is one reason to stay confused and talk things over for a while before etching boards.
Many engineers dislike the confused phase. They want to lock down specs and architectures as soon as possible and enter the implement phase. That is often the path to an inferior product.
Somebody claimed that 19 out of 20 newly introduced breakfast cereals are failures. Most restaurants and dot.com startups fail. PT Cruiser is gone. And there was the carefully-researched, focus-group-tested New Coke.
I thought the low-voltage polymer caps with antiparallel diodes, in the primary, was a cute idea. There is a lot of hostility to ideas here; it's hard to brainstorm in public.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Periodically, when I'm not happy with the way a design is going, I scare the crap out of customers by scrapping the design, and starting completely over.
But they always end up happy!
[snip] ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I'm looking for work... see my website.
Thinking outside the box...producing elegant & economic solutions.
Of course, but they can be useful for quick prototyping. Also the IRS2092 is a rather advanced design, so it's not likely there are counterfeit versions.
Lots of people will continue on a year-til-done design, even if you show them a better thing that can be done in three months.
Some potential customers are turned off by expressions of uncertainty; they want absolute confidence and pretty PowerPoint presentations. Not our kind of customer. The good ones, you can brainstorm and get goofy with and invent things together.
Dare to be goofy!
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at
formatting link
| 1962 |
I'm looking for work... see my website.
Thinking outside the box...producing elegant & economic solutions.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.