Low Voltage, High Current Source for Chocolate Stamping System. 0.8V 10A.

[rantonides]

Hi.  I have an application where we are stamping chocolates with a stainless steel mesh to create a pattern on top.  We want to heat the mesh slightly (45 deg C) to prevent chocolate buildup.  I found that by passing 0.8V DC at about 8A across the mesh, we get just the right amount of heating.  I'm using my adjustable bench supply.  The resistance of the mesh is just over 100 milliohms.

I need to make this a more elegant system.  What is the best way to get down to that voltage?  I thought about changing to AC and getting a stepdown transformer from 120VAC to 0.8VAC, but can't find anything off the shelf at that current.

Off the shelf industrial DC power supplies generally just go down at 5V.

I will want to PWM the power output to control the amount of power transferred.  This will be open-loop.  There's no temperature feedback.

Any ideas?

[andy1]

You don't really need the voltage to be 0.8V that just happens to be the voltage drop over the mesh at 8A (100mohm*8A) so any kind of constant current source should do you fine and I guess you can easily find something for high current LED driving etc.

[rantonides]

I searched Digikey for a constant current LED source with a minimum of 8A.  It seems that the output voltage on all matches can only go down to 9VDC in constant current mode.

[tom66]

You could also power the mesh at approximately 8% duty cycle from a 10V DC source, using something like a 555 timer and MOSFET.

Note the current through the mesh will be approximately 12.5x nominal (i.e. 100 amps) but the average current will be less (8 amps or so.)

This might be too high for any reasonable power supply so you could select a lower voltage like 2.5V. Or use an L-C filter network and a high PWM frequency to limit the peak current.

You could also use a resistor, e.g. for a 5V supply with 0.8V across element and 8 amps current, size resistor for 4.2V/8amps = 0.52 ohms. (0.51 ohms is a close standard value.) The resistor will dissipate I^2 * R heat = 32 watts so it will need to be a large device with a big heatsink.

[oldway]

Variac feeding the transformer of an 12V /10A battery charger. (and a meter to measure the current)

[ConKbot]

Just put 2 120 -> 10V transformers in series. The second transformer should have a 10A+ secondary, so its going to be a 100+ VA transformer, the first can be a tiny little 10-20 VA runt.  The first transformer will divide it to 10VAC (obviously) then it will get divided by 12 again down to 0.8333 VAC

And if the first transformer has spare capacity, you can run control circuitry off of it too.

edit:
Just for an example, first transformer:
http://www.digikey.com/product-detail/en/VPS20-1250/237-1264-ND/666150

second transformer:
http://www.digikey.com/product-detail/en/A41-130-20/595-1286-ND/953156

$56 total (Maybe more if you throw in an enclosure for it and a fuse holder and fuse...) Not bad, imo.

[NiHaoMike]

Take the transformer out of a cheap soldering gun and use a dimmer (or ceiling fan speed control) to fine tune the power.

[Marco]

I will want to PWM the power output to control the amount of power transferred.  This will be open-loop.  There's no temperature feedback.

A synchronous buck converter will do the PWM for you ... bit more elegant than throwing huge chunks of 60 Hz transformer at it and then rectifying/pwm'ing the output.

PS. I just did a digikey search for board mount DC DC converters with output currents between 8A-40A and it seems to me you're spoiled for choice. This was the first result (7$ for a single).

[Niklas]

We did a low voltage, variable high AC current stimuli application a couple of years ago that was pretty simple. A variac followed by a 15 VA toroid transformer with a new secondary winding made of ordinary cable. In our case the output current was measured and any adjustment was made to the variac. I think we used 7 turns of 0.5 mm^2 wound on the outside of the toroid and fixed with some tape to get up to 6 Amps. The wire cross section and number of turns must be matched and trimmed with all the connector and load resistances. Dedicated high current connectors or soldered connections are essential as crappy connectors can ruin the whole setup by adding milliohms. Also consider line voltage variations and effects of heating in copper wire and connectors.

[rantonides]

Take the transformer out of a cheap soldering gun and use a dimmer (or ceiling fan speed control) to fine tune the power.

Just opened up my cheap Yihua soldering/hot air rework station.  Soldering heater element appears to be about 16 ohm.  Transformer has 10.6 VAC, 34V and 26V output taps.  Looks like much higher voltage than what I need.  That would have been a great option, because it already regulates the temperature with PWM, but this load is 0.1 ohm instead of 16 ohm.

[digsys]

No need to mess around with transformers, just pick up something like this -
www.mean-well.com.au/shop/rs-50-50w-single-output-switching-power-supply/28--rs-50-33.html
38 bucks, approved, stable DC voltage. Adding a linear constant current to the end of it is straight foward, and
there'll be minimal heat. There's no advantage to going lower than 3.0/3.3V as 10A is really not a big deal.
There are zillions of similar off-the-shelf alternatives as well, some EVEN have CC modes built in !!

PS: If you need any more help, I'd need a HUGE box of choclate to sample to errr get the parameters right :-)

[rantonides]

PS. I just did a digikey search for board mount DC DC converters with output currents between 8A-40A and it seems to me you're spoiled for choice. This was the first result (7$ for a single).

That looks like the solution.  Can easily get industrial power supply at 5V-15V.  It would be nice if they had it in an industrial DIN Rail version, but looks good.  Thanks.

[T3sl4co1l]

Second the transformer solution.  Industrial reliability, at industrial simplicity: you can't beat a transformer.  A little pricey (you could spare a few amps on the second transformer I think), but it will literally never fail.  I hope your business lasts a hundred years, because the transformers will in that kind of service!

Tim

[NiHaoMike]

Just opened up my cheap Yihua soldering/hot air rework station.  Soldering heater element appears to be about 16 ohm.  Transformer has 10.6 VAC, 34V and 26V output taps.  Looks like much higher voltage than what I need.  That would have been a great option, because it already regulates the temperature with PWM, but this load is 0.1 ohm instead of 16 ohm.

Something like one of those:
http://www.harborfreight.com/180-watt-industrial-soldering-gun-61170.html#.UyEngVYVtb4

[Richard Head]

I vote for a single 15VA custom transformer with a 1V output and use phase control (light dimmer) on the input. You may need quite a heavy pre-load on the dimmer so it behaves.
It should be super reliable due to its simplicity.

Dick

[oldway]

I wrote about choosing a battery charger transformer and there is a good reason for that.

They have a high short circuit voltage to help regulating the secundary current.
The primary and secondary are side by side and not one upon the other to maximize the inductive leakage.
Short circuit voltage is an important caracteristics of a transformer: this show how much is the value of the virtual serial inductance.
In a battery charger transformer, short circuit voltage is between 15 and 20 % of nominal voltage.
In a toroidal transformer, this value is far lower...between 5 and 8 %.

How can we measure this short circuit voltage ?
You short circuits the secondary, feed the primary with a low voltage from a variac and increase this voltage to reach the secondary nominal current.
The primary voltage is a % of the nominal primary voltage: that's the short circuit voltage of the transformer.

Why is this important for a battery charger or for this application ?
With a 15% short circuit voltage transformer 110V/10A, you will need a little more than 16.5V to have 10A in the secundary.
This reduce secundary current variations when secundary resistance vary or when primary voltage does vary.
So it work a little as a current regulation ... this is why its used in battery charger so charging current does not vary too much.

NB: using 2 cascaded transformers is also a good solution because the resulting short circuit voltage is the sum of the short circuit voltages of the two transformers.
Don't use toroîdal transformers, that's the worst choice.

[muvideo]

I would grab a toroidal transformer, wind manually
a low voltage secondary with thick wire, few turns
of 2x2.5mm² pvc coated stranded wire
(the one used for electrical wiring) are enough for 0.8V 10A.

Cheap 35 to 50VA toroidals can be found in halogen
lamps, the "heavy" ones with 12V lamp inside.
For power control slap a triac dimmer for inductive loads
on the primary.

[rantonides]

PS. I just did a digikey search for board mount DC DC converters with output currents between 8A-40A and it seems to me you're spoiled for choice. This was the first result (7$ for a single).

Those DC-DC converters have a remote on/off feature.  I have a query to their tech support to find out if I could use that to pulse the supply every 2 seconds or so to control heat.  I'll keep you posted.  There's not really that much choice at that low a voltage (0.8VDC) and the efficiency really goes to shi*.

[muvideo]

...
For power control slap a triac dimmer for inductive loads
on the primary.

Also, if you want to build the partialization circuit
yourself, you can use burst mode control, using a
normal triac and an optotriac with zero-crossing
detector like a moc3032.
The pulse train will be a slow PWM, you can use
any circuit to generate it, since will be isolated by
the optotriac.

The efficiency will be reasonably high.

[Marco]

I have a query to their tech support to find out if I could use that to pulse the supply every 2 seconds or so to control heat.

Can't you just use a pot to set the output voltage? At 0.59 V you have half the power than at 0.8V ... doesn't that give you enough range of control? (You can probably use an opamp in the feedback loop to get the supply to go below 0.59V too, but you seem to want a mostly turn key solution.)

[dumle29]

0.8v and 10 amps is only 8w. Not really a lot of power. Any current limiting circuit (with that amp capability) should do, and it should drop the voltage low enough to keep the amps? If you can't find anything that will handle 10A without being ridiculously over speced, you should be able to build a simple chopper circuit with a sense resistor, dual op-amp, where the second op-amp is configured as a comparator. You amplify the small voltage over the sense resistor (this needs to be a SMALL resistor, as you only have ~0.8v), and feed the amplified signal into the non-inverting input of the "comparator" and the center tap of a pot in to the inverting input. You can then set the max current (the voltage that will be present over the sense resistor x the gain of the op-amp) and the output of the "comparator" should then run the gate of a mosfet. An IRLB3813 is a high amp mosfet I can think off off of the top of my head. (N-channel, mount it on the low side of the load) It's logic level too.

Now EEV community. Bash everything I'm wrong about. My brain says this should work. Am I spewing garbage?

[dumle29]

@rantonides

So I wanted to do a board, and came up with this schematic in eagle, and laid it out on a small 7.35$ for three boards layout.

I have not tested this, and I have not done anything like it before, but my brain is telling me it should work. Brains are dumb sometimes tho

It uses one SO-8 4081 and gate (kinda a waste, I'm only using one) and one LM324 (using two of those internal OP-amps)

Some support components, and I think it should work.

For R2-5 I'd say go for 1/2W .1% resistors. As per the schematic, 0.1R

If you go for it, I'll MSPaint a hookup diagram.

Schematic:

OSHPark upload: http://oshpark.com/shared_projects/i2vadt6Z

So yeah, any questions?

EDIT: TerminalJack505 played with the circuit in a sim, and caught the second op-amp being backwards, and pointed out the need for a cap. See next page with the updated links

[dumle29]

Rantonides: Did you solve the issue? You got me interested

[David_AVD]

So I wanted to do a board, and came up with this schematic in eagle, and laid it out on a small 7.35$ for three boards layout.

Can you explain how this circuit works please?

[jmole]

I've been using these modules for an upcoming project. Super cheap, and very easy to use.

http://www.digikey.com/catalog/en/partgroup/micro-dlynx-series/563