Yes and those SOT223 surface mount transistors will only dissipate around 2 watts. Supplied directly by DC your standard circuit can be made pretty small even with 2 x TO220...it just seems pretty wasteful to shunt 300mA when the circuit consumes 10mA, and I question whether the circuit would function well with CCS set at, say, 50mA. Hence my thought process for a smaller build.
You can't beat high spare current goodness in shunts as much as you can't beat high bias goodness in class A amps. The big proper Reflektor will always be just that. The proper one. It can carry a Teflon C1 even beyond much spare current and any big C2 you like. Some miniature will be ehh, just good enough.
SOT223 have 55-75C/W junction to ambient. With 1W in a 40C box they are already compromised for lifetime expectancy. With 2W they are nearly kaput. This is constant current i.e. breathless mode we use here. Takes no prisoners.
SOT223 have 55-75C/W junction to ambient. With 1W in a 40C box they are already compromised for lifetime expectancy. With 2W they are nearly kaput. This is constant current i.e. breathless mode we use here. Takes no prisoners.
hi, sorry if this has been discussed already (there's a lot to go through, I'm getting there) but could you recommend parts for the transformer and rectification for the 24v simplistic Shunt regulator. I'm planning to try it in a custom built passive stereo EQ with jfet amplification (2sk170) current per channel is 40ma design based on Hamptone jfet preamp. I'm assuming 2x 15vac secondaries wired series, some diodes but what VAC rating do you guys recommend for the Tx and about the cap and resistor? Good types? Values? Thanks and regards.
Last edited:
No.
VA = output AC voltage * output DC current * 4, so that the transformer runs @ ~50% of capacity to keep it reasonably cool.
If you use *2, the transformer runs @ ~100% of capacity and it will run hot, especially when mains voltage is at it's highest.
eg. for 1 channel 80mAdc continuous from a 30Vac transformer use:
VA = Vac * Idc * 4 = 30 * 0.08 * 4 = 9.6VA
2 channels ~20VA
if you use 30Vac for 24Vdc the regulator drop will be ~45V-24V = 21V
The dissipation will be 21V * (0.08+?) ? = the current that does not pass to the shunt regulator nor to the output. This can vary from 0mA to 20mA
Using 8mA, the dissipated power will be 21*0.88 = 1.85W per channel
Might work without a heatsink, but probably better to use a 30C/W clip on, or bolt to the side/floor.
VA = output AC voltage * output DC current * 4, so that the transformer runs @ ~50% of capacity to keep it reasonably cool.
If you use *2, the transformer runs @ ~100% of capacity and it will run hot, especially when mains voltage is at it's highest.
eg. for 1 channel 80mAdc continuous from a 30Vac transformer use:
VA = Vac * Idc * 4 = 30 * 0.08 * 4 = 9.6VA
2 channels ~20VA
if you use 30Vac for 24Vdc the regulator drop will be ~45V-24V = 21V
The dissipation will be 21V * (0.08+?) ? = the current that does not pass to the shunt regulator nor to the output. This can vary from 0mA to 20mA
Using 8mA, the dissipated power will be 21*0.88 = 1.85W per channel
Might work without a heatsink, but probably better to use a 30C/W clip on, or bolt to the side/floor.
Last edited:
Thanks gents! Can I ask if there is a bom for the simplistic version with r6 mod. I have the bjt's and a few hundred jfets 2sk170 and 117 so should be covered but curious about wattages for resistors, I'm assuming R1 should be 2-5w but are the others okay a 1/4w metal film or should they be higher wattage? Thanks again
Use the SSLV1.1 BIB PDF document and Excel calc that you can download from the builds and fairy tales thread. Its the wrapped up successor of those earlier circuits you look in. Use nominal dissipation of components that is 2-3 times more than what actually is dissipated constantly so to ensure good long term reliability. In the particular reg type this applies to the current setting resistor, all others can be 0.25W.
the *2 or *4 factor takes into account 100% or 50%.
One does need to use *8 for 25% of rating.
But this over-rating of the transformer size will not break it.
Okay thanks Andrew, my understanding was that oversizing the Tx might cause more voltage on the secondary as it isn't loaded heavily, is this true? And how much "wiggle room" from the 30v rectified input is there before the regulation strays from 24v. The only reason I ask is that the jfet stages I've built have been designed and matched per channel to .1% with source and drain resistors for 24v. Likelihood is it would be okay a bit over 24v DC but not a long way
Oversizing will correspond to better regulation under light loading---if your circuitry is depending on a voltage drop due to transformer loading, the transformer is marginally small.
In the good (very) old days, equipment on the mains was not drawing most of the current on peaks of the sinusoidal mains voltage. Now after so many devices are pulling currents on peaks, the peaks get flattened. In a reaction to this, more elaborate circuitry uses power factor correction, so we may be slowly returning to more sinusoidal power.
But the problem is that some power supplies were designed that presumed the voltage was more sinusoidal, so the d.c. generated may be dangerously close to dipping below the minimum required for downstream stuff to stay within regulation.
The use of a given a.c. rating transformer to produce d.c. of the same voltage is sometimes a bit of overkill, but may as well be done for diy. The combination of low-line in brownout conditions and extreme peak flattening can do one in if you're too close to the limits. Remember as well that copper losses go up with temperature.
In a very complex spectrometer I did the electronics for many years ago, there was a fairly-high-voltage supply that in normal laboratory conditions performed flawlessly. Unfortunately at a particular telescope, and often associated with auxiliary equipment cycling on at particular times of the night, the high voltage began to drop out of regulation, and the periodic dropouts created havoc with the preamplifiers, producing huge apparent signals we deemed "spikes". It was a trifle embarrassing to discover that it was just due to very low a.c. line voltage, and the easiest fix was a stepup variac for the particular subsystem.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.