EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: TheresNoPlaceLikeOhm on August 22, 2022, 01:57:29 pm
-
Hi all :)
I'm in need of a means to control the current entering a large electrolysis tank from 1 to 10A. The power supply is a good quality 12V/15A SMPS. Current control doesn't have to be continuous, in 1A steps would be fine.
I've spent many hours in various corners of the internet (here and elsewhere) and have come across various options. Big resistors are the cheapest, but are really inefficient. I'm presently thinking of paralleling 10 seperate 1A sources, and turning them individually on or off as necessary with suitable toggle switches. The devices under consideration are as follows:
MOSFET-based (see attached schematic [attach=1])
3-terminal regulator (eg LT1084 for lower dropout than LM317)
DC-DC CC module ($2 Aliexpress type of thing)
CC LED driver module (more expensive than DC-DC module but better quality, removes need for seperate SMPS)
My personal preference is for the MOSFET-based option, as this will allow me to switch them on/off with an Arduino, plus I get to build it myself.
My concern is this: how will the various outputs interact with one another into a shared load? I'm sure the LT1084 option would need low-value output resistors on each device for current sharing. Would the same be needed for MOSFET modules? And I don't see the switching outputs of the DC-DC CC or LED driver modules playing nicely with one another.
Thank you for any wisdom you might impart ;D
Edit: changing the main SMPS to a higher voltage to handle dropout or buck modules is fine
-
How about just getting a buck converter like a DPS3012 ($30):
https://www.aliexpress.com/item/3256803891297742.html (https://www.aliexpress.com/item/3256803891297742.html)
There's also the DPS5020 and DPS5015. They do a lot more than you need but you also might find other uses for them in the lab.
Also, people have been happy with the more expensive RD60xx line - check out this megathread:
https://www.eevblog.com/forum/testgear/ruideng-riden-rd6006-dc-power-supply/msg4302538/#msg4302538 (https://www.eevblog.com/forum/testgear/ruideng-riden-rd6006-dc-power-supply/msg4302538/#msg4302538)
-
The basic current limiter in your schematic will get hot. But this goes for readymade linear current sources/ LM317's and the like. Calculate the worst-case V*I across the MOSFET or regulator.
You can probably parallel LM317s configured as individual sources. Each source has its own control loop. Sacrifice a bag of 317s and a bag of eBay mini toggle switches to science.
Attached are examples of classic voltage or current controlled precision sources. Simple to make using Darlington’s or MOSFETS. Fairly precise (not Marco precise) but usually good enough.
Check the pass transistors can handle the worst your power supply can chuck at it. Heatsinking a must.
The terminal voltages on DC boost and mains LED current sources can rise to a dangerous voltage without a load attached. You have been warned.
-
Big resistors are the cheapest, but are really inefficient.
They are no more inefficient than your MOSFETs or 3-terminal regulators. For a linear solution the excess power has to go somewhere. You'll find no difference in efficiency between those three options.
An interesting parameter missing in your post is the voltage compliance needed. Is 10...12 V really enough?
-
Using transistors instead of resistors will not matter for efficiency of a linear design.
To get high efficiency, a switching buck converter could be modified for constant current operation, which involves adding feedback from a current shunt to control the output.
-
If you dont mind manual regulation one option is to put a Variac in front of your low voltage transformer.
In case the plan is to build everything from scratch a 10 or 20 amp transformer form an old battery charger would do the job. It dont matter the secondary voltage either 12V or 24V will do the job.
The Variac dont have to be very large at 12V, 10A is only 120W or at 24V 240W so if you get something between 250VA and 500VA you should be OK.
Of instrumantation you will need a amp meter. If you can get one with a transformer you are able to have it inline without a powersupply. Look for a meter with 100A scale and a transformer 100/5A. By wrapping the transformer with 10 turns you will get 100A on the meter at 10A.
100A meters and 100/5A transformers are probebly the most common ones.
-
At one place I worked we used DC welders as power supplies for electropolishing.
-
Hi all :)
I'm in need of a means to control the current entering a large electrolysis tank from 1 to 10A. The power supply is a good quality 12V/15A SMPS. Current control doesn't have to be continuous, in 1A steps would be fine.
I've spent many hours in various corners of the internet (here and elsewhere) and have come across various options. Big resistors are the cheapest, but are really inefficient. I'm presently thinking of paralleling 10 seperate 1A sources, and turning them individually on or off as necessary with suitable toggle switches. The devices under consideration are as follows:
MOSFET-based (see attached schematic (Attachment Link) )
3-terminal regulator (eg LT1084 for lower dropout than LM317)
DC-DC CC module ($2 Aliexpress type of thing)
CC LED driver module (more expensive than DC-DC module but better quality, removes need for seperate SMPS)
My personal preference is for the MOSFET-based option, as this will allow me to switch them on/off with an Arduino, plus I get to build it myself.
My concern is this: how will the various outputs interact with one another into a shared load? I'm sure the LT1084 option would need low-value output resistors on each device for current sharing. Would the same be needed for MOSFET modules? And I don't see the switching outputs of the DC-DC CC or LED driver modules playing nicely with one another.
Thank you for any wisdom you might impart ;D
Edit: changing the main SMPS to a higher voltage to handle dropout or buck modules is fine
Hello there,
Post power supply current limiters are always a problem. They dissipate heat.
The worst thing in power conversion is handling the power more than once. That means more than one regulator instead of just one.
The right way to do this is similar to the other thread here about CC regulation. The idea is to tap into the original power supply feedback node and control that with a small amount of additional circuitry. That's not always possible though and that may be the case with your setup too. The other thread is the one:
"Adding Constant Current CC Feature To Power Supply".
There are a few ways to do this without heating up the whole house at the same time (much wasted energy).
1. Use a second power converter, namely, a buck converter that can handle the entire load current. That would be a converter that has a current set adjustment.
2. Use multiple buck converters, maybe 1 amp each, and switch them in or out as you were thinking. Might also have to use small value current sharing resistors also because some converters do not like to be directly connected in parallel.
3. Get a completely new power supply with current adjustment. The cost of some of these came down quite a bit over the past 10 years or so. You could probably find one that does 10 amps at 30 volts for around $60 USD. Shipping costs depend on where you get it.
For #1 it isnt that bad of an option. You would get current adjustment with that although it may be a trim pot you may have to replace yourself with a regular pot if you intend to adjust a lot.
For #2 you can find buck converters that do 1 amp for probably something like $1 each on Amazon. They often sell several in one pack.
#3 of course would be the most expensive but you'd have everything you need in one box with no extra wiring laying around.
-
Thank you for the replies, everyone :)
How about just getting a buck converter like a DPS3012 ($30):
That's a good option, thank you kindly!
Using transistors instead of resistors will not matter for efficiency of a linear design.
They are no more inefficient than your MOSFETs or 3-terminal regulators.
Very good points. I hadn't thoroughly thought through the mechanism of action until you pointed this out.
The basic current limiter in your schematic will get hot. But this goes for readymade linear current sources/ LM317's and the like. Calculate the worst-case V*I across the MOSFET or regulator.
You can probably parallel LM317s configured as individual sources. Each source has its own control loop. Sacrifice a bag of 317s and a bag of eBay mini toggle switches to science.
Attached are examples of classic voltage or current controlled precision sources. Simple to make using Darlington’s or MOSFETS. Fairly precise (not Marco precise) but usually good enough.
Check the pass transistors can handle the worst your power supply can chuck at it. Heatsinking a must.
In the case of MOSFETS / 317s I would mount the array of TO220s on a big honkin' heatsink. Regardless of final choice I reckon some scientific sacrifices should be made, even if just for the learning ;D. Thank you for the circuits!
If you dont mind manual regulation one option is to put a Variac in front of your low voltage transformer.
I'd prefer to stay away from mains voltage. A well-made and properly certified DC source removes a lot of undue worry and potential kabooms.
Hello there,
Post power supply current limiters are always a problem. They dissipate heat.
The worst thing in power conversion is handling the power more than once. That means more than one regulator instead of just one.
The right way to do this is similar to the other thread here about CC regulation. The idea is to tap into the original power supply feedback node and control that with a small amount of additional circuitry. That's not always possible though and that may be the case with your setup too. The other thread is the one:
"Adding Constant Current CC Feature To Power Supply".
There are a few ways to do this without heating up the whole house at the same time (much wasted energy).
1. Use a second power converter, namely, a buck converter that can handle the entire load current. That would be a converter that has a current set adjustment.
2. Use multiple buck converters, maybe 1 amp each, and switch them in or out as you were thinking. Might also have to use small value current sharing resistors also because some converters do not like to be directly connected in parallel.
3. Get a completely new power supply with current adjustment. The cost of some of these came down quite a bit over the past 10 years or so. You could probably find one that does 10 amps at 30 volts for around $60 USD. Shipping costs depend on where you get it.
For #1 it isnt that bad of an option. You would get current adjustment with that although it may be a trim pot you may have to replace yourself with a regular pot if you intend to adjust a lot.
For #2 you can find buck converters that do 1 amp for probably something like $1 each on Amazon. They often sell several in one pack.
#3 of course would be the most expensive but you'd have everything you need in one box with no extra wiring laying around.
Useful information, for which I am grateful :).