Linear Technology DC2132A CV&CC Adj. Bench Power Supply Board

[BravoV]

Spotted at Linear Technology Journal, the DC2132A Adjustable bench psu board released in May 2014, I think size wise is quite nice, and also it has switching pre-regulator with linear front end of course.



Efficiency wise stated in their spec sheet with Vin = 36 Volt & Vout = 3.3 Volt @ 3 Amp is at 60%, not bad at all isn't it ? While with Vin = 12 Volt & Vout = 5 Volt @ 3 Amp is reaching 70% with approx < 7 Watt of power loss.

All details like the schematic, manual, performance tests like transient responses (overload included) and etc with claimed ripple < 20 mV at 60 uF Cout.

Details of DC2132A product page here -> Linear Technology DC2132A - 24V 3A Constant Voltage, Constant Current Bench Supply

or

directly at the details & performance page (PDF) -> High Performance Portable DC Bench PS : Save Money and Free Up Bench Real Estate by Building Your Own

Price ? $200 a pop. 

Obviously they had to use all LT's made chips  , its just really like to hear our resident experts here about this circuit design or a discussion about it.

[DutchGert]

Looks very nice.

I have used the lower power LT8614 in a design and that one really preforms well.

If u add two simple voltage displays to the output and Imon u can even monitor your settings. Very nice.

Add a transformer (or a black box input), a output on/off/set switch and u can make a nice little kit. Let me now if u need someone to design the board .

[nowlan]

I would say spin up some boards, $200 too rich. I think the soldering is not very fun for a kit.

I always wonder why no one has cloned the older HP power supplies.

[DutchGert]

I would say spin up some boards, $200 too rich. I think the soldering is not very fun for a kit.

I always wonder why no one has cloned the older HP power supplies.

True, I meant a kit as in a custom assembled pcb with a housing etc.

[timb]

Hmm, this looks nice. I'm thinking it would be easy to get uC control by adding a dual OpAmp and dual DAC. Use that in place of the pots on SET and ILM.

In the schematic they show an INTVCC rail but don't provide a source for it. I guess you need to have a seprrate converter to power everything?


Sent from my Smartphone

[DutchGert]

Hmm, this looks nice. I'm thinking it would be easy to get uC control by adding a dual OpAmp and dual DAC. Use that in place of the pots on SET and ILM.

In the schematic they show an INTVCC rail but don't provide a source for it. I guess you need to have a seprrate converter to power everything?


Sent from my Smartphone

IntVCC is generated from the LT8614

[timb]

Hmm, this looks nice. I'm thinking it would be easy to get uC control by adding a dual OpAmp and dual DAC. Use that in place of the pots on SET and ILM.

In the schematic they show an INTVCC rail but don't provide a source for it. I guess you need to have a seprrate converter to power everything?


Sent from my Smartphone

IntVCC is generated from the LT8614

Ah, okay, good catch!


Sent from my Smartphone

[timb]

Those LT3081's look pretty neat!

It looks like the Vset pin sources 50uA of current and is typically set by a single resistor.

So I'm trying to figure out the best way to put this under MCU control. The issue I see is that this current + resistor is used to set a voltage on the non-inverting input of the internal error amplifier. It seems this set voltage is equivalent to the desired output voltage, I.e. If you want 5V, you'd use a 100k resistor (50uA * 100k = 5V).

If you had a current DAC that could sink or source current, I'm thinking you could set it up so that when it's supplying 0uA the output voltage would be 12V (based on the regulators internal 50uA source, so you'd use a 240k resistor), then you'd use the DAC to inject or absorb more or less current to change the voltage from 0 to 24 volts.

It's late and I'm most likely over thinking this, I'm sure. Would love to hear ideas though. 


Sent from my Smartphone

[pelud]

If you had a current DAC that could sink or source current, I'm thinking you could set it up so that when it's supplying 0uA the output voltage would be 12V (based on the regulators internal 50uA source, so you'd use a 240k resistor), then you'd use the DAC to inject or absorb more or less current to change the voltage from 0 to 24 volts.

I'm not sure how well it would work, but what about just having an op-amp drive a transistor with the voltage at the set point as the feedback?

[timb]

Interesting. Where is the 1.2V, 50V and 12V coming from?

[pelud]

Interesting. Where is the 1.2V, 50V and 12V coming from?

The 50V would be the unregulated input going to the LT3081, the 1.2V is an example set-point voltage coming from the micro (via a DAC or maybe PWM + RC filter?), and the 12V is the voltage going to the SET pin.  If you load the .txt file into the falstad circuit simulator you can play around with the set-point via the voltage slider on the right.  The 50uA current source is of course the one built into the regulator between the input and the SET pin.

[rdl]

I'm not saying this isn't a nice board and I'm probably wrong about this, but it seems like it's way over-priced. 

I see just 5 ICs and a handful of passive components, and the ICs are made by the same company selling the board. They sure aren't paying retail price for them.

I've bought a number of other dev boards and such of equal or higher component quantity for 1/3 to 1/2 the price of this one, and I'm not talking about from China either. Obviously I must be missing something here.

[dannyf]

Nothing unusual.

The switching regulator is configured to drop a constant voltage over the regulators, vs. the ground. I mentioned this as one efficient way to deal with heat dissipation in one of the power supply threads.

[timb]

It’s basically a Dev Board, and like all Dev Boards (TI’s LaunchPad’s aside), there’s an arbitrarily high cost associated with it. I’ve found you can get these type of things for free if you just ask a lot of times.

Anyway, the benefit here is that the schematic and BOM is available, so recreating it won’t be hard. I wish Gerbers or CAD files were provided though. (That’s one thing I *LOVE* about TI. They provide CAD, Gerber, PDFs, Source Code, pretty much everything for a large number of their boards.)

Edit: Wait, here we go: http://www.linear.com/docs/44898

Full Schematics, PCB, Gerbers… Nice!

[timb]

Interesting. Where is the 1.2V, 50V and 12V coming from?

The 50V would be the unregulated input going to the LT3081, the 1.2V is an example set-point voltage coming from the micro (via a DAC or maybe PWM + RC filter?), and the 12V is the voltage going to the SET pin.  If you load the .txt file into the falstad circuit simulator you can play around with the set-point via the voltage slider on the right.  The 50uA current source is of course the one built into the regulator between the input and the SET pin.

Ah I see now. The image was loading at low resolution for me (it’s fine now) so it made it hard to see exactly what was going on.

Seems like a very sane approach. I’ll do some more advanced simulations with it on the full circuit.

The other issue I see for a bench top version is that if you’re simply using a transformer and rectifier on the front end to power it (120VAC to 36VAC @ 3A), that’s going to be a pretty big transformer I imagine…

I wonder how well this would run with a boost converter in place of the buck. Might be a good option for a battery powered portable supply.

[timb]

The other issue I see for a bench top version is that if you’re simply using a transformer and rectifier on the front end to power it (120VAC to 36VAC @ 3A), that’s going to be a pretty big transformer I imagine…

I wonder how well this would run with a boost converter in place of the buck. Might be a good option for a battery powered portable supply.

Digikey has some transformers

http://www.digikey.ca/product-detail/en/VPS24-5400/237-1277-ND/666163

Thats about the cheapest suitable one. You would probably be better off finding a laptop brick, that will likely be cheaper then a transformer,bridge,high rms current Capacitor, fuse a line filter/TVS ETC. If you skim their documentation they in fact recommend a laptop brick.

Yeah, after my post I started looking. At $35 for that transformer, you’d be better off just getting one of these.

It’s hard to find laptop bricks at that voltage/current rating it seems like. I did manage to find one, but it’s only 36V@2A.

Something else they mention in the manual is that you can add a third linear regulator to bump the output to 4.5A; but I’m finding that when you get past 150W or so with those Chinese AC to DC supplies, they start getting very big and have fans, which defeats the purpose entirely.

Anyway, that 120W supply I linked above looks like the ticket. I’d have to change the capacitors out and verify the clearance was okay, but it’s not putting out a tremendous amount of power, so I don’t think *that much* can go wrong, so long as it’s safe.

[dannyf]

I’m finding that when you get past 150W or so with those Chinese AC to DC supplies, they start getting very big and have fans, which defeats the purpose entirely.

That's why having the pre-regulator referenced to the output is so incredibly important: in this case, the voltage the linear regulators are dropping is just 1.7v, regardless of the output voltage.

Jung has the same concept in his regulators.

[timb]

I’m finding that when you get past 150W or so with those Chinese AC to DC supplies, they start getting very big and have fans, which defeats the purpose entirely.

That's why having the pre-regulator referenced to the output is so incredibly important: in this case, the voltage the linear regulators are dropping is just 1.7v, regardless of the output voltage.

Jung has the same concept in his regulators.

Either way though, *something* has to power the pre-regulator. Now, what might interesting is to use the AC/DC SMPS itself as the pre-regulator. Normally they use a TL431 or other shunt based reference to do the regulation, so you’d just have to inject in to that and you could stay in regulation and 1.7V above your output.

The issue I can see with this, however, is that fact that most of these AC/DC SMPS designs are using an isolation transfer with a turn ratio that’s specific to the output (obviously). As you start lowering the voltage, the output side rectification diodes are going to be blocking more and more energy, which means they’re going to heat up pretty quickly. You’ll also be lowering your duty cycle, which means a bigger output capacitor will be required, which means worse load regulation… You also might have flux issues in your transformer at very low duty cycles. Synchronous rectification could solve some of this. Combine that with a multi-tapped transformer and a very smart SMPS controller and it could work I suppose.

[DutchGert]

These boards are never meant to go into big production. Just a way for Linear to show of there components.

Btw, they not always cost a lot, after showing interest in one to my local Linear rep he just told me one is coming my way free of charge. I love Linear for that.

[BravoV]

These boards are never meant to go into big production. Just a way for Linear to show of there components.

Btw, they not always cost a lot, after showing interest in one to my local Linear rep he just told me one is coming my way free of charge. I love Linear for that.

Yep, this is intended as sales & promotional item for their customer for free, I guess LT deliberately put that kind of price tag just to piss off potential retail buyers.

Please do some torture review once you have it in your hand.

[DutchGert]

These boards are never meant to go into big production. Just a way for Linear to show of there components.

Btw, they not always cost a lot, after showing interest in one to my local Linear rep he just told me one is coming my way free of charge. I love Linear for that.

Yep, this is intended as sales & promotional item for their customer for free, I guess LT deliberately put that kind of price tag just to piss off potential retail buyers.

Please do some torture review once you have it in your hand.

Will do

[DutchGert]

I am thinking of brewing up a schematic combining this with some proper SET and OUTPUT ON/OFF circuitry and combining it with a ideal diode bridge (LT4320), a nice torroid transformer ( 80VA 2x12V Transformer ) and some proper input protection to make a nice 'Green' linear PSU (as far as u can call a linear psu green ofc). 

[kt315]

 I was quite curious about this schematic as well. I have no doubt LT does not pay nearly close to $200 for the boards, but ...

I loaded the BOM into digi-key. There are quite a few costly components on the list aside from the LT parts + 4 layer board. I was in fact quite surprised by the cost of some passives they are using.

I probably will try to build one of these for entertaining/educational purposes, I'd be surprised if it would be less than $100 in cost for single quantities.
The switcher chip seem to be PITA to solder (at least at my level of experience).

[DutchGert]

I was quite curious about this schematic as well. I have no doubt LT does not pay nearly close to $200 for the boards, but ...

I loaded the BOM into digi-key. There are quite a few costly components on the list aside from the LT parts + 4 layer board. I was in fact quite surprised by the cost of some passives they are using.

I probably will try to build one of these for entertaining/educational purposes, I'd be surprised if it would be less than $100 in cost for single quantities.
The switcher chip seem to be PITA to solder (at least at my level of experience).

True, they use quite a bit of odd-valued/rare components. It clearly was design't as a show-off of there parts.

That said, I like the design as they present it so perhaps it can be made into something rela nice. Especialy if u add another 3081 so u have 4,5A output which makes it more useful in the lower voltage ranges.

[DutchGert]

It took a bit longer than expected but I had a go at testing the board.

So far it performs really nice.

Power consumption
Idle load @ 35VIN, no output connected = 17mA

PowerIn @ 35VIN, 3V3 3A Out = 35V x 470mA = 16,45W
PowerOut @ 35VIN, 3V3 3A Out  = 3,3V x 3A = 9,9W
Ploss = 16,45W - 9,9W = 6,55W
n = ~60%

Temperature LT3080 @ 25C Amb
@ 35VIN, 3V3 3A Out:
Temp 1 = 0,63V
Temp 2 = 0,65V

Load regulation (on output terminals)
@ 35VIN, 3V3 3A Out :
Vout no load = 3,30V
Vout full load = 3,28V

Noise and transient response
Noise no load: ~10mV
Noise 3A load:  ~28mV

0-3A Transient response overshoot: ~160mV
3-0A Transient response undershoot: ~316mV