Reflektor-D builds

Jump to Latest
#483
Hi Salas,
sorry for late response but the work .. :-(

Salas said:
If the max current it takes is bellow 600mA then use two wire output mode by jumpering 0S to 0F and S+ to F+ locally at the reg's output connector to see if it continues dropping.
Click to expand...

That was it! 😱 Mmmmh but why? The connection between WaveIO and DAC was only 8cm. Ok, i'll try it again with shielded cables in the final enclosure (and report here).

Thanks again!
 
#484
Don't know specifically in this case, there should be some devices that can prove difficult to this reg for phase margin or for noises they couple back. Its a very wide bandwidth regulator. Nice thing is there are wiring options.
 
#487
Ok, 8V target >800mA gives? 6,4ohms? Why ship a 20ohm dummy for up to 5V and recommend 30ohm for 7V?

The only answer I can think of is because these values are maybe the most common currentvise at these voltages 🙂. I need to stress mine though 😉

Regards
 
Last edited:
#496
bern said:
Hi,
Please advise me a R1 value for 580mA receiver load draw.
Will a 0.9 Ohm be OK?
The table from p.2 of the user guide shows R1 values only up to max 450mA.
Tia.
B
Click to expand...

0.9R will produce about 680mA current limit by the 0.61/0.9 formula

When the "receiver's" max demand is a verified 580mA then 680mA limit is enough

Such an arrangement will produce little heat on the output MOSFET sink with the most heat dissipated from the CCS MOSFET sink.
Only 100mA * Vout for M2. It will burn the 680mA - 580ma difference times 5V in other words. But full 680mA * (Vin-Vout) for M1.
Because the sinks are grouped together, the heat should average between them to some extent
 
#497
Turbon said:
Ok, 8V target >800mA gives? 6,4ohms? Why ship a 20ohm dummy for up to 5V and recommend 30ohm for 7V?

The only answer I can think of is because these values are maybe the most common currentvise at these voltages 🙂. I need to stress mine though 😉

Regards
Click to expand...

The dummy load's purpose is to represent a lifelike enough ratio of consumed power versus available constant power for the final connection to your application. So everyone can prepare his REF-D safely before his valued DAC or USB interface board etc. is connected. Nobody bets good on a premier without rehearsal. Possibly disastrous mistakes could happen to digital chips mainly in the voltage setting procedure while choosing LEDS etc. before any chance to realize and correct.

The 20 Ohm "universal" dummy just cuts a midway path between the look up table CCS and load examples in the manual for up to 5V output. Tries to be helpful for most likely digital voltage rail uses when must be having one part in hand and make not many calculations. Its a Swiss army knife, not a custom tool beyond a popular frame of applications.

In your case the voltage and current demands are already unusual for the intended purposes of the reg. By considering the standard build heat exchange capabilities that is. The circuit itself does not reject scaling up otherwise. So calculate a dummy that represents your final load in mA at rated output voltage. NOT a dummy that represent your CCS limit at Vout. That would equalize limit & demand and bring it to CC mode, i.e. cutting voltage down and allowing no further mA, just like the lab supplies do for protection. Leave at least 100mA headroom for CCS setting.

Beyond being a good thing having this current source for the better regulator performance, it is also a built in protection indeed if something goes wrong in your DAC etc. It does not let your device pull crazy power from the REF-D on any fault, effectively doing damage control, or it signals some bad modding like you tried a new part or a tweak that is done wrong etc. Saves most trouble in the system and the regulator itself.

If you will reach much heat like having 65C+ on the sinks, better use external ones by wiring M1 & M2 with short wires and insulating their tabs. Another way it could be using a 60mm 12V DC fan blowing on the board level sinks. Better sitting on their top. If such a fan is powered up to 8V DC it usually goes unnoticed. There are special "silent" models also. Not really, but clearly better than most others.
 
#499
That headroom is what goes up as heat regarding M2 mostly. It brings down modulation THD in the rail from the load if its a rich headroom. More noticeable if analog circuits or analog chip portions are involved. Another way to guard against such effects if a tonal thickening is suspected is to increase the internal gain by the already known R6 mod to JFET replacement.
 
#500
Right Salas. Thanks for the ideas about 8V but it ain't a Salas circuit without leds 😀...

An externally hosted image should be here but it was not working when we last tested it.


No current yet - R1 2ohm and dummy 27ohm. But voltage in ballpark 8.17V and temp on sinks 29,5C after an hour measured deep deep down in the belly between the sinks 🙂 ambient - 25,6 due to all machinery creating heat. Lowest dc in to maintain 8.17v is 9.67VDC but I will use 10.64V´ish from an 12V transformer as you prescribed. Oh yes - I visited 15VDC in area - still 8.17V'ish.

Current next 🙂


I heard or read rumours about an R6 - fet mod? Why and how?

Thanks and regards
 
Last edited:
You must log in or register to reply here.
Top Bottom