A couple of small isolated voltage converters with bipolar outputs
I'll start with why I bought them.
While preparing the platform for a new electronic load, I decided it would be better to power the analog section from a bipolar supply after all. That immediately avoids the problems you get with very low voltages across the current-sense shunts when measuring small currents. And to avoid being tied to a common ground, I chose isolated converters. I also liked the price.
I actually ordered three types of converters, but the third type doesn't fit the title because it has neither galvanic isolation nor a bipolar output (although it does have two voltages). I'll tell you about it next time.
Everything came packed in the usual clear bags.

First up is the Tyco ME005BK. I bought a pair; each one costs about $1.33–1.4 (it was $1.25 when I bought them) - link.

This is a branded converter, so there's a full datasheet. In short:
Input voltage - 48 V
Full input voltage range - 39.5–60 V
Output voltage - ±12 V
Output current - 210 mA
As you can see, unlike the converters I showed earlier, this model is designed to run from a single “nominal” voltage—48 V—so all the tests will be done at that voltage.

The converter is housed in a small case, with soft potting compound filling the bottom.

Dimensional drawing.

These converters are (or were) made with various output voltages, in both single- and bipolar-output versions. My version is highlighted.

There won't be too many tests; I'll try to keep things relatively brief.
First, I checked its overload capability, with the full output connected, i.e. 24 V.
Although it's rated for 210 mA, the converter easily delivers 300 mA, but at higher current it goes into protection. The voltage stays very stable, only sagging under overload.

No-load power consumption is about 0.45 W. The readings below are at 100, 150, 210, 250 and 300 mA, respectively.

And here it is as a table.

Another test checks voltage regulation with a 100% unbalanced load: one rail is fully loaded while the other has no load.
1, 2. Measuring the positive rail while loading the negative one; on the left, no load.
3, 4. Measuring the negative rail while loading the positive one.
In the first case, the loaded rail was at 11.64 V; in the second, 11.11 V.
In my opinion, that's a pretty good result, especially since I don't plan to use it in this mode.

A few oscilloscope traces.
1. 12 V, no load
2. 24 V, no load
3. 12 V, full load
4. 24 V, full load.
The traces for the negative and positive rails are identical, so I'm showing only the positive rail.
I wouldn't call the result good, but it's still below the datasheet specification.

It gets warm in operation. The left photo was taken after about 10 minutes, the right one another 20 minutes later.

The second module is made by Martek power. It's three times more powerful and rated for a wider input voltage range of 36–75 V, but it has the same ±12 V output. It currently costs about $0.9 (it was $1.56 when I ordered it) - link.
It looks quite a lot like the previous converter, but this one doesn't have a sealed case; it has ventilation holes. Then again, it's three times more powerful.

General specifications from the datasheet.

The converter is available with three input voltage options—12, 24 and 48 V—and three output voltage options—±5, 12 and 15 V—for a total of 9 models.
In my case, the model has a 48 V input and ±12 V output.

Module dimensions and pin layout.
There are also two versions, with and without an enable function. I have the full version, i.e. with a control input.

Despite having three times the power, it’s almost the same size as the previous one.

Taking it apart revealed that

A closer look.


The control input works differently from the converters I’ve shown before.
1. The input is active-low; apply voltage to turn it on.
2. The converter still works when the control input is disconnected, but I connected it to the positive supply anyway, just to be safe.
1, 2. Standby current draw at 36 and 62 V
3, 4. No-load current draw in operating mode at 36 and 62 V.

Although the rated output current is 630 mA, it handles up to 800 mA without a problem. Beyond that, it goes into protection, but up to that point it works perfectly.

But when loading just one output, it doesn’t work quite so nicely.
1, 2. Measuring the positive rail while loading the negative one; the voltage on the left is the no-load voltage. The loaded rail dropped to 10.67 V, while the unloaded one rose to 13.44 V.
3, 4. Measuring the negative rail while loading the positive one. This is where the problem showed up: at a load current of 500–600 mA, the converter goes into protection. At 400 mA it works, but the unloaded rail rises above 16 V while the loaded one drops to 7.9 V.
It’s clear these converters were probably designed for an even load, but the imbalance here is rather extreme.

Current draw at different load levels (25, 50 and 100%) and input voltages of 36 and 60 V.

Efficiency shown as a table and graph.

Since it wasn’t very convenient to load each channel separately here, I measured the total output voltage. These are the waveforms at 0, 33, 6 and 100% load, with an input voltage of 48 V.
I also connected 100 µF, 25 V Samwha RD capacitors across the outputs, since the module description said external capacitors were required.
There’s a pronounced spike, although the overall ripple isn’t that large. I think a simple LC filter would improve things considerably, though it may have been caused by a ground loop with the electronic load.
The datasheet specifies ripple of up to 60 mV, but as I understand it, that’s at 12 V (one rail).

The module’s temperature under full load, measured 10 minutes after startup and again 17 minutes later. The maximum temperature reached 85 °C. The input voltage was 48 V.

I can’t say whether these are used units or just old stock: the first datasheet dates back to 2008, and the units themselves have no manufacturing date. For the price, I think they’re pretty good. They have plenty of headroom in terms of load capacity. The ripple could be lower, so I’ll have to add output filters, though I’d have done that anyway. Otherwise, I’m happy with them.
The first module weighs 24 g, the second 16 g, so the second ones are probably a better deal both in terms of price-to-power ratio and shipping cost.
That’s all from me. I hope you found the review useful.
https://www.kirich.blog/obzory/preobrazovateli/703-para-nebolshih-preobrazovateley-napryazheniya-s-galvanicheskoy-razvyazkoy-i-dvupolyarnym-vyhodom, dc-dc преобразователь с гальванической развязкой своими руками, схема attiny13 импульсного преобразователь получение двухполярного напряжения 3,7 вольт, dc/dc преобразователь 12/12 с гальванической развязкой схема, понижающий dc dc преобразователь с гальванической развязкой, устройствоcn6009 как получить двухполярный выход, двухполярный ключ с гальванической, двуполярный dc dc стабилизатор, преобразователь с двуполярным выходными напряжениями / 12 в, маломощный преобразователь разнополярного напряжения, dc-dc преобразователь с гальванической развязкой, схема конвертора напряжения с 9 в двухполярное 12 вольт, преобразователь 1,5 в в двухполярное питание, dc dc 12 5v с гальванической развязкой своими руками, dc dc 9 вольт в двухполярное 9 в, 12в в двухполярное автомобиль, схема маломощный преобразователь с гольванической развяской12 на 5 ватт своими руками, преобразователь dc/dc с гальванической развязкой своими руками схема, схема слаботочного преобразователя с 9 в крона на двуполярное 12 вольт, hef4093 и lm2903 в импульсном преобразователе напряжения из однополярного в двухполярное, понижающий преобразователь с двуполярным выходом, схема dc dc двухполярного преобразователя, преобразователи с двуполярным выходом самодельные, понижающий dc-dc преобразователь с гальванической развязкой своими руками, lm2577 гальваническая развязка, обзор преобразователя напряжения, технический обзор, преобразователь напряжения, преобразователь с гальванической развязкой
Questions and answers
How much current can the Tyco ME005BK ±12 V converter handle?
Although rated for 210 mA, it ran up to 300 mA before entering protection.
How does the Martek power converter perform with an unbalanced load?
When only one output rail is loaded, the voltages diverge considerably: in the test, the loaded rail fell to 7.9 V while the unloaded rail rose above 16 V. At full load, the module reached 85 °C.
Which converter offers better power-to-weight value?
The author considers the Martek power module the better value: it is about three times more powerful, weighs 16 g versus 24 g for the Tyco, and handled up to 800 mA in the test.
The figures from this review are in the measurement database: Converters
Discounted products
You may be interested in
Discounted products
All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!
Comments1
Only registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.
Log inRegister
Comments are published in their original language.