
Home  Forums  Rules  Articles  The diyAudio Store  Gallery  Blogs  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 
Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum 

Please consider donating to help us continue to serve you.
Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving 

Thread Tools  Search this Thread 
Yesterday, 07:07 PM  #1 
Previously known as kingden
Join Date: Aug 2008
Location: Evanston, IL

Transformer Sizing, etc
In my scanned notes is a basic working of how my curve tracer will operate. I finally found time to get back to it .
I performed some rough calculations trying to figure out the average power dissipation of the B+ supply. I chose for a calculation benchmark a CCS load of 0.5 amps. The unregulated B+ rail puts out around 700V into the PAD195 and IRFPG50 pass mosfet. B is 100V. Since the B+ supply puts out a ramp, it will dissipate a constantly changing number of watts depending upon the time. I programmed the device so the ramp takes 2 milliseconds to reach 580 volts (B+ max). Therefore, when the ramp hits 50 volts, the instantaneous dissipation will be 325 watts. At 580 volts, we have 60 watts. I do not know the method to calculate average power dissipation for a 2ms ramp offset by 100ms. Therefore, I assumed the B+ supply is a square wave putting out 50V, making the regulator dissipate 325W for 2ms. Therefore, I found the average power dissipation to be 7 watts as the duty cycle is 2%. 2ms is a pretty short time, but those numbers still seem overwhelming for that mosfet. The mosfet datasheet says that amount of time, given the range of voltages and the 0.5A benchmark I am within the SOA. http://www.mouser.com/ds/2/427/91254104428.pdf I have a feeling I am missing something in this calculation (I am thinking average power dissipation), so I would appreciate it if someone could possibly set me straight. Also, how does one size a transformer for the B+ and B rail for this application? The unregulated B+ has 47uF worth of capacitance across the rail. B has 22uF. Last edited by BRSHiFi; Yesterday at 07:09 PM. 
Today, 12:22 AM  #2 
diyAudio Member

This isn't that hard to figure out, just some basic calculus (attached). The voltage is 46.89V_{RMS}; call that 47V_{RMS} as design nominal voltage.
Since this waveform stays above the centerline at all times, there's also an average of: V_{ave}= 5.67V, as a DC component. As far as what you've got for power, that depends on the load resistance. As for the PTX specs, that depends on the actual load, the current demand, and whether you decide on a Cinput or Linput ripple filter. The Cinput makes better use of PTX voltage, while Linput makes better usage of current carrying ability. If you're expecting 0.5A_{P} at 580V, that's: R= 580/0.5= 1160R, which gives: 47^{2}/1160= 1.9W. The average current will be: 47/1160= 40.52mA  the value you'd use to design the DC rail supply. That you have such a low duty cycle makes these values so small. The one thing to remember is that the DC rail capacitor needs to be large enough to supply the peak current demand, which you said is 0.5A. 0.5A * 47= 23.5W 23.5 * 2.0E3= 47E3 (newtonmeters) W= 0.5CV^{2} C= 42.55uF (47uf design nominal) Last edited by Miles Prower; Today at 12:38 AM. 
Today, 05:30 AM  #3 
Previously known as kingden
Join Date: Aug 2008
Location: Evanston, IL

I did the algebra and paired the equation down to the attached image.
The B+ regulator will dissipate 23.5 Wrms with that waveform running into the load continuously. Assuming 15 curves are drawn into the 0.5 amp CCS load, the waveform will only run for about 1.6 seconds. That is not enough time for the system to come up to temperature. Also, a tube is not a CCS load. I picked the 0.5 amp as just a reference. So assuming the 0.5A load, what heat sinking do you recommend? I have one on there that can adequately dissipate about 10 watts. 
Today, 06:08 AM  #4 
diyAudio Member

Since your current isn't going to be anywhere close to 0.5A for most of the time, and you'll probably never see that anyway (unless you're curve tracing a big RF power or TV HD type) then heatsinking for 10W should be more than adequate.
It's the behaviour when that ramp drops that I'd be concerned about. A high speed Schotkey reverse biased across the MOSFET would be a good idea, if this diode isn't already included on the die itself (check the spec sheet). Also, a series high speed Schotkey diode in series with the drain of the MOSFET is also a good idea. This should protect against any inductive flyback due to stray inductance. Sometimes, MOSFETs fail due to excessive dV/dt problems, and nothing can prevent it since this happens on die, and nothing external can protect the gate against that. You could consider a slight modification that ramps down instead of dropping suddenly. Say, ramp down over 0.2mS. 
Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
output transformer sizing  is there such a thing as to big?  JZatopa  Tubes / Valves  56  27th May 2013 10:46 AM 
Sizing Isolation Transformer  Lavcat  Everything Else  5  24th January 2012 01:20 AM 
Output transformer sizing...again...  rsumperl  Tubes / Valves  17  2nd September 2010 04:14 PM 
Sizing capacitors  preiter  Power Supplies  5  29th May 2006 05:48 AM 
Output Transformer Sizing?  DrStrangelove  Tubes / Valves  6  14th November 2004 08:25 AM 
New To Site?  Need Help? 