Re: Power supply project

[Jinx-4]

Re the 30V 3A variable V and I supply

http://www.electronics-lab.com/projects/power/003/

I've tracked down the amendments

http://home.clear.net.nz/pages/joecolquitt/30V_mods_notes.gif

http://home.clear.net.nz/pages/joecolquitt/30V_mods1.gif

http://home.clear.net.nz/pages/joecolquitt/30V_mods2.gif

http://home.clear.net.nz/pages/joecolquitt/30V_mods3.gif

http://home.clear.net.nz/pages/joecolquitt/30V_mods4.gif

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Dennis Crawley-2]

On Sunday, July 13, 2008 10:05 AM [GMT-3=CET],
Jinx  wrote:

> Re the 30V 3A variable V and I supply
>
> http://www.electronics-lab.com/projects/power/003/
>
> I've tracked down the amendments
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods_notes.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods1.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods2.gif

This link appears to be missing...

>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods3.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods4.gif


Thank you for post your findings, Jinx. I like the Re resistor, I think with
5W is enough according to your maximum current (3A)...
BR,
Dennis


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[dpharris]

Dennis Crawley wrote:

> On Sunday, July 13, 2008 10:05 AM [GMT-3=CET],
> Jinx  wrote:
>
>  
>> Re the 30V 3A variable V and I supply
>>
>> http://www.electronics-lab.com/projects/power/003/
>>
>> I've tracked down the amendments
>>
>> http://home.clear.net.nz/pages/joecolquitt/30V_mods_notes.gif
>>
>> http://home.clear.net.nz/pages/joecolquitt/30V_mods1.gif
>>
>> http://home.clear.net.nz/pages/joecolquitt/30V_mods2.gif
>>    
>
> This link appears to be missing...
>
>  
>> http://home.clear.net.nz/pages/joecolquitt/30V_mods3.gif
>>
>> http://home.clear.net.nz/pages/joecolquitt/30V_mods4.gif
>>    
>
>
> Thank you for post your findings, Jinx. I like the Re resistor, I think with
> 5W is enough according to your maximum current (3A)...
> BR,
> Dennis
>
>
>  

Yes, thanks for that Jinx.  What's the next step?  Are pcbs still
available?  Maybe we should add a PIC and V&A monitoring LCD?

David



--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> http://home.clear.net.nz/pages/joecolquitt/30V_mods2.gif
>    
> This link appears to be missing...

Sorry about that boss. Should be there now

> Yes, thanks for that Jinx.  What's the next step?  Are pcbs still
> available?  Maybe we should add a PIC and V&A monitoring
> LCD?

That's the intention

Electronics Lab suggest this, using a TQFP ATMega (a bit OTT)

http://www.electronics-lab.com/projects/test/022/index.html

http://www.electronics-lab.com/projects/test/022/schematic.jpg

I'm undecided what display to use. I want to have several PSUs in
one box, possibly 4 or 5, so an LCD each seems a little extravagant.
OTOH, I've a box of 4-digit 1/2" displays and drivers, and plenty of
7-segment LEDs (even a few old 1/8" calculator displays !!!)

First things first - get one of these beasties made up and see how
it performs

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

BTW, in

http://www.electronics-lab.com/projects/test/022/index.html

the author is making a voltage measurement across the current sense
resistor to get the 'current flowing' display. It might be desirable to
take a measurement at the current-limiting pot also and display that
as well. And/or an LED to show that current out >= limit set

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Dennis Crawley-2]

On Sunday, July 13, 2008 5:59 PM [GMT-3=CET],
Jinx  wrote:

> I'm undecided what display to use. I want to have several PSUs in
> one box, possibly 4 or 5, so an LCD each seems a little extravagant.
> OTOH, I've a box of 4-digit 1/2" displays and drivers, and plenty of
> 7-segment LEDs (even a few old 1/8" calculator displays !!!)

I'll use the 7 segments. When things are going wrong I want to clearly see
the
Amps or the Volts. (summer project)

> First things first - get one of these beasties made up and see how
> it performs

I have the negative side also. In the same cabinet I've installed a PC power
supply with all the outputs in the front pannel 3.3,±5,±12, One 9V output
for the PicstartPlus :)

BR,
Dennis



--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vasile Surducan-2]

On 7/13/08, Jinx <joecolquitt@...> wrote:

> Re the 30V 3A variable V and I supply
>
> http://www.electronics-lab.com/projects/power/003/
>
> I've tracked down the amendments
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods_notes.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods1.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods2.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods3.gif
>
> http://home.clear.net.nz/pages/joecolquitt/30V_mods4.gif

Assuming we need someday 3.3V and 3A from this power supply, the
thermal dissipation on Q4 and Q5 will be around 90-100W. I think from
this perspective that is a pour design. Imagine how would be at +30C
ambient and four hours of running. Even power supplies designed back
in 1980 have two good solutions for this problem:
1. switchable trafo windings using relays
2. a controlled bridge using two thyristors and two diodes (sometimes
four thyristors).
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> Assuming we need someday 3.3V and 3A from this power supply,
> the thermal dissipation on Q4 and Q5 will be around 90-100W

That is a significant disadvantage of linear PSUs. For a drop like that
you'd expect to need fan cooling. I have a product that needs 4.1V @
3A for extended periods and for that I use an LM2576 switcher, which
of course hardly heats up at all. So you do use what's appropriate when
necessary

One of my musty-smelling old books is the GE SCR Manual and they
go into switched bridges in some depth

One reason for this project though is that I'd like to build and share
something that could be put together from spare parts that most
would have or could get cheaply, even if it does double as a heater
sometimes

Rummaging around through many years' worth of hoarding has
turned up some useful finds. Parts that I'd completely forgotten were
there. LM338, 2N3055, MJ15003 and even MJ11016 power
Darlingtons, all unused. Even really old LM309 in steel TO3. I
think I inherited a lot from a friend who closed his repair shop many
years ago

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[William "Chops" Westfield]

>> Assuming we need someday 3.3V and 3A from this power supply,
>> the thermal dissipation on Q4 and Q5 will be around 90-100W

I've had two commercial "lab" linear power supplies, and they both  
seem to engage relays (audible clicks) at set voltage points along  
their range.  The 50V supply that's working now clicks at about 15  
and 30V...

BillW

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> I've had two commercial "lab" linear power supplies, and they
> both seem to engage relays (audible clicks) at set voltage points
> along their range.  The 50V supply that's working now clicks at
> about 15 and 30V...

What do you think they would be switching ?

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Marcel Birthelmer]

My guess is they have a few secondary windings from the mains transformer
that get rectified to a few different DC voltages, and going between output
ranges switches between the input signals. That way, you the output
regulator doesn't have to deal with dissipating I x 29V when you want a 1V
output, but instead just up to the next stage.
- Marcel

On Wed, Jul 16, 2008 at 5:47 PM, Jinx <joecolquitt@...> wrote:

> > I've had two commercial "lab" linear power supplies, and they
> > both seem to engage relays (audible clicks) at set voltage points
> > along their range.  The 50V supply that's working now clicks at
> > about 15 and 30V...
>
> What do you think they would be switching ?
>
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> My guess is they have a few secondary windings from the mains
> transformer

Hmmm. There's a thought. One transformer I'd probably use has
secondary taps. I guess for using the control circuit I've posted you
could add a boost switcher so that the control circuit always has a
constant V+, no matter what DC is going through the main bridge to
the output transistors, unless the PSU has a small secondary just for
the electronics. LED displays are pretty hungry so it actually would
be a good idea to have a separate supply for them rather than waste
capacity of the main winding. And changing the secondary would
certainly go some way to reducing losses at low output voltages

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

The guess is correct. That is pretty standard in bench supplies.

Sean


On Wed, Jul 16, 2008 at 10:39 PM, Jinx <joecolquitt@...> wrote:

>> My guess is they have a few secondary windings from the mains
>> transformer
>
> Hmmm. There's a thought. One transformer I'd probably use has
> secondary taps. I guess for using the control circuit I've posted you
> could add a boost switcher so that the control circuit always has a
> constant V+, no matter what DC is going through the main bridge to
> the output transistors, unless the PSU has a small secondary just for
> the electronics. LED displays are pretty hungry so it actually would
> be a good idea to have a separate supply for them rather than waste
> capacity of the main winding. And changing the secondary would
> certainly go some way to reducing losses at low output voltages
>
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Martin Klingensmith]

I wonder if it would be a good design to have a buck pre-regulator SMPS
that was heavily filtered followed by a linear regulator dropping a few
volts. The buck supply would have a slow feedback response because of
the filtering but the floating linear reg could make up for it.
-
Martin
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

That sounds like a decent idea, although it may still have some of the
disadvantages of a switcher (switchers are blamed for being noisy, but
often this noise is very high frequency which a linear regulator
cannot reject very well. So, the way to fix it is to design the SMPS
correctly in the first place and add shielding/filtering.

Sean


On Fri, Jul 18, 2008 at 11:06 AM, Martin <martin@...> wrote:

> I wonder if it would be a good design to have a buck pre-regulator SMPS
> that was heavily filtered followed by a linear regulator dropping a few
> volts. The buck supply would have a slow feedback response because of
> the filtering but the floating linear reg could make up for it.
> -
> Martin
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> I wonder if it would be a good design to have a buck pre-regulator
> SMPS that was heavily filtered followed by a linear regulator dropping
> a few volts. The buck supply would have a slow feedback response
> because of the filtering but the floating linear reg could make up for it

Seems to me that, like the relays which switch secondaries, at certain
voltage points you could change the SMPS's feedback resistor ratio
to minimise the voltage drop across the following linear section (if you
have one)

How would you effect adjustable current-limiting in a simple SMPS ?

> switchers are blamed for being noisy, but often this noise is very high
> frequency

Sean, a modem I use specifies a very low noise tolerance so I looked
at SMPS ripple output. AFAICT it was pretty low, under mV at least,
and I didn't take any special precautions to remove it. A schematic
straight out of the datasheet actually. A filter (cap and small series
inductor) is a common 'dashed line' option in many d/s schematics

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[John La Rooy]

On Sat, Jul 19, 2008 at 10:28 AM, Jinx <joecolquitt@...> wrote:

>
> Seems to me that, like the relays which switch secondaries, at certain
> voltage points you could change the SMPS's feedback resistor ratio
> to minimise the voltage drop across the following linear section (if you
> have one)
>
> Could you not just add a zener in the feedback leg of the SMPS so it always
tries to output Vzener volts above the voltage you have dialed?

Then you always have 3V (or 5V or whatever you choose) headroom for the
linear to operate.

John
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jinx-4]

> Could you not just add a zener in the feedback leg of the SMPS
> so it always tries to output Vzener volts above the voltage you have
> dialed?

Hmmm, so it's tracking but somewhat higher than the required output

I'll look into that. The feedback divider reduces the output voltage
to that of the controller's internal reference, and the controller
switching works to maintain the output voltage by comparing that
feedback voltage to the reference

Possibly a TL431 adjustable Zener might be used ? If voltage setting
is to be done in the linear stage then that information has to go back
into the SMPS section for tracking. Perhaps

Time to open up some app notes ......

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Apptech]

>> Could you not just add a zener in the feedback leg of the
>> SMPS
>> so it always tries to output Vzener volts above the
>> voltage you have
>> dialed?

> Hmmm, so it's tracking but somewhat higher than the
> required output

At the minimum the tracking "preregulator" needs to be a
Vdropout above the linear output.

I used an arrangement a few years ago which I saw also used
recently in a circuit from AFAIR Olin. I had a very nice LDO
linear regulator  with a dropout of a few tenths of a volt.
I placed a transistor across the regulator (say NPN for the
following example) with emitter to ouput and base via  a
resistor to input. Collector via a resistor to a higher
voltage. When across-regulator voltage exceeds Vbe the
transistor turns on and pulls the collector down to Vout. Or
use a PNP and reverese base+R and emitter connections and
collector will pull high from ground when regulator drop is
 > 1 x Vbe. Use this to control smps pre-regulator. Worked
very well..
In this case the smps was a CD40106 hex Schmitt trigger
driving an NPN switch to produce 5 volts from 4 x C cells.
When batteries were new the batteries fed via the inductor
to the linear regulaor and the smps was gated off. As
batteries aged *OR* when the local motor load loaded the
batteries below about 5.2V the smps ran. So the logic always
saw 5V. Off state quiescent current was well under 100 uA
including 2 x LDO following.

The regulator was an AIC1722. Taiwanese sourced. Better than
most and very low cost. (??? $US0.10 in volume)

Data sheet
http://others.servebeer.com/misc/ds_aic1722.pdf



    Russell


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist