
Home  Forums  Rules  Articles  The diyAudio Store  Gallery  Blogs  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 

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 
1st October 2012, 10:21 AM  #1351  
diyAudio Member
Join Date: May 2007

Quote:


1st October 2012, 10:27 AM  #1352 
diyAudio Moderator
Join Date: May 2003
Location: Palatiw, Pasig City

secondary resistance reffered to primary = Rsec + [Rpri(Es/Ep)^2]
__________________
http://www.elab.ph/forum/index.php?topic=32688.0 
1st October 2012, 11:37 AM  #1353 
diyAudio Member
Join Date: Mar 2008

oops. thats primary resistance referred to the secondary.
as long as the transformer magnetising current is well behaved (basically sinusoidal) and fairly small, then reflecting the primary series impedance to the secondary will be a good approximation. If, however, the transformer pushes close to (or into) saturation, then the mag current will be extremely "peaky" and the voltage drop across the primary series impedance will be a lot higher than expected. I believe the technical term for this type of transformer, in audio applications, is "lousy". and note the use of the term "series impedance"  leakage inductance is usually the dominant parameter in transformer regulation. look at the transformer parameters Tom has been using. The normalised impedances are: Rp = 2.8%, Lp_leak = 9.4% Rs = 5.2%, Ls_leak = 9.5% for both the primary and secondary circuits, leakage inductance dominates. at rated current the voltage drop across each impedance is its PerUnit impedance times the rated voltage. the resistive and reactive voltdrops add vectorially. so the primary has sqrt(2.8^2 + 9.4^2) = 9.8% total voltdrop  the primary resistance is completely swamped by the primary leakage inductance. yeah Rp dissipates watts, but Lp_leak does the damage to voltage regulation and the secondary has sqrt(5.2^2 + 9.5^2) = 10.8% total voltdrop  the secondary resistance is again swamped by the secondary leakage inductance. it gets even worse, because the rectifiercapacitor filter draws nonsinusoidal current with a high crest factor  IOW lots of odd harmonics. resistance goes up with the square root of frequency due to skin effect, but inductive impedance is directly proportional to frequency  so yet again the leakage inductance swamps the resistance. thats quite a few parameters. but using Tonys suggestion and making the above assumption of a nonsaturating transformer, we should specify the transformers secondaryreferred leakage inductance and resistance in addition to turns ratio & VA. And Andrew is quite right  I would expect the transformer resistances to vary widely between different devices. In addition I would expect even larger variations in leakage inductance  which depends entirely on the actual transformer construction. Years ago I learned the hard way to very carefully specify wound components. we had always designed our own SMPS magnetics, but the Ac line choke designs were farmed out to a magnetics manufacturer  we specified dimensions, inductance, operating current and that was about it. All fine and dandy, until our manufacturer burned to the ground. The new place met our "specs" but as we'd left off all the important information (core sizes, no. of turns, peak flux density that sort of thing) the resultant parts were useless. Last edited by Terry Given; 1st October 2012 at 11:46 AM. 
1st October 2012, 12:07 PM  #1354 
diyAudio Member
Join Date: May 2007

With significant inductance it could be a bit more complicated than just a voltage drop. The inductance will try to keep current flowing so the charging pulse could be lengthened on the trailing side, although slowed down on the leading edge.

1st October 2012, 12:27 PM  #1355 
diyAudio Member
Join Date: Mar 2008

DF96,
yeah it does. and it spreads out the conduction angle. this drops the crest factor quite a bit  instead of a CF around 5..7, it'll be more like 3..4. Power Factor gets better too. Schaeffers analysis shows this in detail, complete with pretty pictures (hint hint) when designing AC line chokes for offline rectifiercapacitor filters, there is a critical inductance below which there isnt much useful effect. its about 3.5%  which is a bit below the typical leakage inductance range. its helpful enough that it occurred to me at the beginning of this thread that the transformers should be designed with L_leak = 5% (which is a lot better than Toms transformer) but no smaller. 
1st October 2012, 09:03 PM  #1356 
diyAudio Member

Terry,
I did get a whole new transformer model, baed on AndrewT's measurements of a 250 VA secondary, which I am now using instead of the original model. I will try to remember to post the new model parameters after I get home. Or you could look back at what AndrewT posted. The new model seems to have much better regulation. Tom Last edited by gootee; 1st October 2012 at 09:06 PM. 
1st October 2012, 09:15 PM  #1357  
diyAudio Member

Quote:
Don't worry too much, yet. Changing the capacitance totaly changes the picture. But changing the transformer output voltge changes it even more. I purposely chose a power level and capacitance combination that made sweeping the VA rating cause significant changes, with the given 28 VRMS transformer output voltage. With a little more excess transformer output voltage, and any largeenough capacitance, the circuit becomes much less sensitive to everything else, especially the VA rating. Having a highenough PSU transformer VRMS Out is extremely important, for use with a power amplifier. Tom 

3rd October 2012, 07:10 PM  #1358 
diyAudio Member

Minimum Reservoir Capacitance Calculator (xls)
Attached is a betatest version of an Excel spreadsheet that calculates the minimum required reservoir capacitance, given the line frequency, desired maximum RMS output power, transformer RMS output voltage, nominal load resistance, and the minimum voltage across the amplifier between the power supply and the load.
It uses the calculations shown in post 1339, which were first outlined in post 1246. Those posts are at Power Supply Resevoir Size and Power Supply Resevoir Size . I did not insert any errorchecking. If an impossible combination of transformer output voltage, nominal load resistance, and desired maximum output power is entered, then the calculated capacitance values will be negative. Total rectifier voltage drop is currently "hard coded" as 1.4 Volts. But it could easily be changed. Transformer VA rating is not yet taken into account. Cheers, Tom Last edited by gootee; 3rd October 2012 at 07:25 PM. 
4th October 2012, 12:05 AM  #1359 
diyAudio Member

In my version of the spreadsheet, I changed the rectifier voltage drop from 1.4 to 1.7 Volts, since that's closer to what it is, in the simulations.
I did a series of runs (see attached plot), with Pout=50W and transformer Vrms=28V, automatically stepping the capacitance (25004500 uF by 500uF) and the VA rating of the transformer (100500 VA by 100 VA, and plotted the "minimum voltage difference between the power rail and the high side of the load" (over 0.2 second) vs the reservoir capacitance and the VA rating. The minimum difference voltage (i.e. the voltage range where the amplifier resides) must stay above about 3 Volts, to avoid distorting the output with the power rail voltage ripple. The plot verifies that the spreadsheet very closely predicted the minimum acceptable reservoir capacitance found by simulation, for the 50W / 8 Ohms / 28 Vrms case. In this case, the calculation appears to have been equivalent to using a 400 VA transformer. The spreadsheet, if accurateenough, could be used to create a table and/or plots that could show, for example, the regions of the [Pout, Vrms] plane where the capacitance values would be "reasonable". I am also hoping to find out how to estimate the effect of the VA rating on the required capacitance, using sets of simulation runs like the ones that were used to create the attached plot. Edit: I added a plot with some of the raw timedomain data from the LTSpice runs, covering about 5% of the total time. For each run, the data for each entire run was automatically examined by the .MEASURE scripts, and the one minimum VrailVload point was found, for each run, and the voltage difference was calculated and saved, then used for the first plot. Cheers, Tom Last edited by gootee; 4th October 2012 at 12:23 AM. 
Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Valve power supply  How to size transformer?  SanderW  Power Supplies  25  4th January 2013 04:12 PM 
How do you calculate choke size in a power supply?  Original Burnedfingers  Tubes / Valves  25  5th January 2012 12:23 AM 
power supply bypass cap size  BigE  Power Supplies  11  5th July 2011 02:59 PM 
Power Supply Case Size  diymixer  Power Supplies  1  10th October 2010 05:47 AM 
What size power supply should I get for repair work?  spooney  Car Audio  3  6th December 2007 11:50 PM 
New To Site?  Need Help? 