[time-nuts] Power supply for Thunderbolt

[Dave Brown]

----- Original Message ----- 
From: "David C. Partridge" <david.partridge at dsl.pipex.com>
To: "'Discussion of precise time and frequency measurement'" 
<time-nuts at febo.com>
Sent: Monday, April 21, 2008 8:40 PM
Subject: [time-nuts] Power supply for Thunderbolt


> I'm in the process of buying a Thunderbolt, and will need to provide 
> it with
> some volts.
>
> I sat down over the weekend to design a linear supply for it. 
> Initially
> things looked OK using a 12-0-12 transformer, and a few IC voltage
> regulators, but then I did some worst case design and dissipation 
> analysis,
> and found that I needed to use a 15-0-15 transformer to allow for 
> a -10%
> mains variation, and add pass transistors and lots of heat sinking 
> to all
> the regulators to allow for a +10% mains variation.
>
> Add to that transformer regulation issues (over-size transformer), 
> the time
> to design the PCB and the cost of getting it made, and things 
> started to
> look expensive :-(.
>
> I can probably source a linear supply on eBay, but these are 
> typically
> pretty large and heavy.  For example: eBay item number 330201077319
>
> So is it reasonable to use an SMPS, and add additional filtering to 
> the
> output?  Typically these talk in terms on 150mV ripple & noise which 
> I need
> to reduce by a factor of ten or better.
>
> If you have been there, done that, I'd be interested to hear what 
> you did.
>
> Thanks
> Dave Partridge
>

----- Original Message ----- 
From: "David C. Partridge" <david.partridge at dsl.pipex.com>
To: "'Discussion of precise time and frequency measurement'"
<time-nuts at febo.com>
Sent: Monday, April 21, 2008 8:40 PM
Subject: [time-nuts] Power supply for Thunderbolt


> I'm in the process of buying a Thunderbolt, and will need to provide
> it with
> some volts.
>
> I sat down over the weekend to design a linear supply for it.
> Initially
> things looked OK using a 12-0-12 transformer, and a few IC voltage
> regulators, but then I did some worst case design and dissipation
> analysis,
> and found that I needed to use a 15-0-15 transformer to allow for
> a -10%
> mains variation, and add pass transistors and lots of heat sinking
> to all
> the regulators to allow for a +10% mains variation.
>
> Add to that transformer regulation issues (over-size transformer),
> the time
> to design the PCB and the cost of getting it made, and things
> started to
> look expensive :-(.
>
> I can probably source a linear supply on eBay, but these are
> typically
> pretty large and heavy.  For example: eBay item number 330201077319
>
> So is it reasonable to use an SMPS, and add additional filtering to
> the
> output?  Typically these talk in terms on 150mV ripple & noise which
> I need
> to reduce by a factor of ten or better.
>
> If you have been there, done that, I'd be interested to hear what
> you did.
>
> Thanks
> Dave Partridge


Hi David- that Meanwell unit looks a good choice.

For the record, I measured  the following current requirements for an
OEM Tbolt using a 428B.

Measured Thunderbolt DC supply rail loadings -

Initial loads-within 10 -20 sec from power applied.

+12  730 mA

+5   ~400 mA

-12 not checked



After 2 minutes-

+12   200 mA (sudden drop from ~700 mA, just prior to 2 mins)

+5  300 mA

-12 not checked



After 3 hours-

+12  150 mA

+5  300 mA

-12  ~ 2 mA

 Tbolt mon running (i.e. serial port activity) in latter two cases-not
for initial load test

 Regards
 DaveB, NZ