negative supply

K sounds like a good challenge will see what I can come up with this week

Yup, point taken. but gosh 20 BUX in small volume. That thing better shit gold!

Reply to
bitrex
Loading thread data ...

Related thing while you're here ;) I have a need for converter for 5- 8 volts input to 12 volts output at about 100 watts, I'd like to buy something off the shelf but I couldn't immediately find something that has specs like that.

I thought about flyback but ugh transformers. WebBench came up with this design using the LM5155 peak current-mode controller when select "optimize efficiency", it's just a straight boost:

It says frequency is 300kHz. Does this look vaguely cromulent? 500V 100 amp MOSFET, 4 electrolytic input caps and 6 electrolytic and 7 plastic output caps, yikes.

Reply to
bitrex

This is a lot to ask of a wimpy sot-23 switching IC. Step up (haha) to a sensible TO-263 package. TI's LM22677 handles up to 42V, 5A and works at 500kHz.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

in the case of my design, the main priority is size, with noise being second most important. cost is almost a non-issue for this board. the LT8643S integrates the FETs in a 4mm package, and it also integrates 0.1uF bypass caps.

Reply to
sea moss

We pay about half that. That's not too bad for a fully packaged switcher. They are reliable and quiet.

The little LM7805 replacement switcher bricks look interesting.

formatting link

They are cheap, and most can convert positive to negative too.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

The 54302 is rated for 3 amps, but that gets really hot. It's fine at

2 amps. We use fat traces on several pins to spread the heat out. They are great when you need a bunch of goofy supplies, like +12 and +5 and +3.3 and +2.5 and +1.8 and +1.2.

I've been talked/bullied into doing the design with no negative supplies, so now I need to boost my +24 to about +30, still around 2 amps maybe. The goal is to use both halves of a TCA0372 in bridge mode, driving a 2:1 step-up transformer, to get maybe 30 or 32 volts RMS to power synchros and resolvers. The PCB will be brick-walled with

12 giant transformers, so maybe we'll put all the electronic parts on the bottom side.

Something like this:

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

You're kidding right? Production methods and labor costs were different, LITERALLY HALF A CENTURY AGO.

But just to humor you -- I don't recall noticing /any/ trimmers inside my TDS460. ;-D

And that's Tek, a company that could afford to do ludicrous, expensive things, at least in their more rarefied, upscale machines.

They wouldn't last long, doing that kind of crap today. Not with Chinese scopes flooding the market, that use the technology from Tek's biggest competitor (Keysight's, uh, what was it called, Wave-something ASIC), at a $400 price point.

I wonder how many DSOs you can buy for the price of just one of those boatanchors (in inflation-adjusted dollars). Of comparable or better analog performance, mind.

Let alone multi-channel, measurements and signal analysis (FT, logic, decoding) features, that you literally couldn't buy for any price (in one single instrument), back in the day.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
 Click to see the full signature
Reply to
Tim Williams

What sort of 22 uF 1206 do you use? Distortion would be bad news in that application.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Tim Williams wrote

Just makes you wonder why everybody hangs on to those good old Teks now does it? You never calibrated a Tek in your life. I have many, quite a procedure, by the book, took a real amount of time, had to be verified by someone else. No chances taken. Sure those were multi channel, sure they had processors, and sure those were 1 G samples / second.

Yes, as I mentioned before, large scale integration will increase speed and make things cheaper with less adjustment. However if you REALLY want to go top of the line, then you are most likely right back with trimmers and calibration procedures. Else your products are at the level of just anyone else.

Most people can now afford to buy a Rigol or whatever and use it to look at circuits. In many cases it gives me the impression of people looking at their sandwich with a microscope. 'But I really need that, I design with micro processors'.

Decent analog scope few MHz bandwidth will do fine for most cases and give more reliable results. hehe Same for your inductive heater.

I designed and build my own scopes, from just a few MHz with tubes, to solid state with RTL logic, Also build a 300 MHz wide analog using a circuit that somebody from Tek published in an electronics magazine. Met that guy later...he told me they were pissed he published that circuit...

If I even needed a scope like that, then it is old hat. Now that was nice stuff. Not that I ever needed it, donated it when I left the country.

It has always been 'the challenge' to do it, for me at least, and then move to something else. Apply in one field what you learned in the other.

Reply to
<698839253X6D445TD

John Larkin wrote

...

I do not see that, but have not been in the US for a few (20?) years, last time I looked LA was full of illegals looking to make a buck, but OK.

Wow, and snow falls for free...

Yes, one thing that drew may attention about a year ago was someone of those weather observers going on about the importance of some effect / airflow (something, am no expert on that) crossing the equator for the first time observed,. There was a whole discussion about the effect that would have. I am counting on N America becoming like Sahara desert, but it could be a new ice-age too. So much for that,. But something in what that guy wrote made me look and I could not see it

formatting link
that site I sample once a week and have hundreds of pictures of wind flow that can be put together into a movie that spans years. One of my hobbies or interests I guess
formatting link
round and round it goes. Nice site.

Reply to
<698839253X6D445TD

John Larkin wrote

Why are you swithing on te high power side, not on the low power side?

Reply to
<698839253X6D445TD

It's often difficult to use digital adjustments in high-Z or high speed circuits. Digital pots and MDACs are great for tweaking offsets but have a lot of capacitance that makes them progressively inaccurate at high frequencies, especially at low gain settings. Little trimpots have a great deal less. A drop of glyptal makes them non-volatile.

You used to be able to get GPIB-controlled screwdrivers for automatic tuning, so it can all be algos anyway. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

What's with the CPC1008N switch? Its Ron is 8 ohms. And those opto FETs switch slowly, in the ms region.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

The switching happens infrequently so that speed doesn't matter. We have that SSR in stock and in the PADS library so I can do the layout with that. Currently, the Triad transformer losses dominate, but if I buy a custom transformer I can reduce losses, so I'd want a better SSR.

The next problem is to not fry anything, or collapse the main power supply, if some loads get shorted. I'm thinking that I can add a low value resistor from the opamp V- to ground, and snoop that current, and kill the DDS dac for that channel, or open the SSR, for a while if we see too much power going into a channel. Maybe have enough capacitance somewhere in the power supply that we don't reboot the uP and FPGA meanwhile.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

It's just a commodity X5R. The nominal operating frequency is 400 Hz. Maybe I should use a big polymer. There should be theoretically zero DC, but I don't want 10s of millivolts DC applied to the transformer.

I do have a 1000 uF 2.5V polymer in stock. That would be better.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

A Tek 545 used to cost as much as a Chevrolet. The ratio is now roughly 100 scopes (much better scopes) per car.

A serious individual could start a growing electronics company in his garage, for a couple thousand dollars.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I wonder how much digital signal processing is done in modern scopes, instead of trim caps and pots.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

No surprise there: the cost of semiconductors goes up fast with peak current, slow with peak voltage, so most power systems (and

100W certainly qualifies) are designed for step-down where possible. If you wanted 48V to 12, 'specs like that' would be thick on the ground.

Well, yeah: add to that there's DC in the inductor (you'll want to wind your own, with a gap) and that catch diode doesn't look easily obtainable either (400mV at 25A? Really?). When I last handled 5V, 20A, without a lot of good semiconductor solutions, it took a transformer, heatsinked rectifiers, etc.

Old-school can do this, though: have you considered an auto generator driven by a DC motor? An auto salvage yard can get you those parts, with current capacity to spare.

Reply to
whit3rd

I came up with a design (on paper at least) that should be able to pull it off. Also there are some prefab modules that are kinda similar spec on eBay (but nothing that fits that requirement exactly) that gave me a general idea of what a design should look like-ish.

it'll be fairly simple basically just a current-mode controller IC at a

100kHz or something, TO-220 MOSFET and Schottkey rectifier on a couple of inch long heatsink, with some fins maybe, about a 10uH inductor wound on a ~1" OD toroid core (the magnetics "Kool Mu" composite type has good power density and BSat of about 0.8T) with about ten turns of a couple "strands" of 10-12 gauge magnet wire in parallel. 7 output caps is probably overkill for me I could probably get away with three.

I'll try that next...;)

Reply to
bitrex

LM5122?

Reply to
Lasse Langwadt Christensen

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.