Re: Adding up all the currents! I'm lost!
Hi,?
?
A couple of suggestions that might be of some help:
?
In PSUD:
?
- When you start up, there will be some large peak currents. You might want to use "after a reporting delay" and pick 1 second to work with the values you have. This will let some of the high peak currents settle down and reduce the RMS to give you some steady state figures
- When adding loads, you need to load the left hand side sections rather than just the right. Insert a current tap just before R1 and load it with 80mA for starters, this might be a good value for a pair of 6V6s. Insert a second current tap before R2, I'm guessing this is for 6V6 screens use 4mA as a starting figure
- Finally, change the 5K at the end for a current load. Try 10mA if it's just for some preamp stuff, more if you have a tuner etc.
- The transformer voltages don't look out of place but the resistance looks low for powering a pair of small output tubes. How did you measure/calculate this??
For your PSU build in practice:
?
- Try and think of ways to avoid the Zeners; while they will drop voltage they will greatly amplify any voltage regulation problems so will be far from ideal
- Your 325-0-325 as you've guessed is probably going to be over the top. Try and source another (hamfests etc.) or swap with someone for a lower voltage, 250-0-250 or 275-0-275
- If you can't source/swap/beg/borrow there are companies like Hammond etc. that make a wide range of new transformers
- Check out schematics of other amplifiers / PSUs that have been made before - it's always a good source of inspiration and a reference point to calibrate your ideas
?
General:
?
- Data on the 6V6 can be found here: , the linked data sheets will give typical voltage and current draw information for a variety of different amplifier classes
- Data on the 5Y3 can be found here: , the linked data sheets show a variety of transformer input voltages and expected DC outputs which you can use to calibrate your work on PSUD
?
Hope this helps, please message back and let us know how you get on or if you get stuck with any of the above.
?
Regards,
Duncan
?
|
Adding up all the currents! I'm lost!
Hi all,
I am attempting to understand this PSU2 app. I want to design a power supply for a tube amp that has a B+ of around 350Vdc, and B+1 at about 330Vdc, and my final B+2 (preamp supply) at about 180Vdc. But I have yet to get it to match on my bench test bed. I think the main issues is the load! and the current being drawn.
I have a PT that can supply about 100mA at 325-0-325. Meaning 650V, then after the rectifier I have about 477Vdc.Which is way too high for my 6V6 tubes!. I then use 3 zener diode of 33V drop to bring down the voltage to about 380Vdc. at my first C1 cap. (meaning that is B+). However, depending on the load I put on the PS, changes this of course, and I am reading in the range of about 300Vdc. Right now as a test, I put a 5K resistor at the end of the line.
?
My question for you all. is in regrades to current draw. On the PSU designer it shows in the chart, the calculated RMS voltage and current. To get my total current being pulled from the transformer do I add up all these currents? If so, then I have way exceeded the max current on this transformer!
?
Additionally, I am totally guessing on how to load down this PS. I stuck that 5K on the end. Its drawing about 30mA at C3. But is that too much? or too little? I know a typical 6V6 push pull amp would put out about 15W at about 125mA. But I am designing this to run at a much lower wattage. Something more like 10W or even 8W if possible.
?
See my simulation below. the voltage at C1, C2, C3. would be acceptable, if the currents were correct and my load correct. But with these component values on my real test bed, I am a bit off. Please note, that since the PSU app does not allow you to add in series diodes, I had to fudge the T1 voltage to compensate for a 99V drop, so you will see it listed at 285V, not the 325V it is speced for.
Please help. Do I have the correct values for R1 and R2, R3?
Thanks.
?
|
Re: Where is “download PSUD2” now?
It seems to run "on many Linuxes"; no mention of Mac or Win.
?
?
Hi Paul, I had a look over at the Appimage site and for sure it's aimed at aligning different Linux distributions rather than different OSes. That in itself might cause problems with the binaries as there are different binaries for 64 bit AMD/Intel, 32 bit Intel, 64 bit ARM (Raspberry Pi, etc.) I need to have a look however, as it appears to solve a problem I have with packaging Linux apps. A job for later in the week :)
?
I am seeing minimal-spec Windows machines under $300
?
I've bought 5 mini PCs since August, two cost real money, however three of them have been bargain basement "Celeron" type machines going for GBP ?149 with 500GB solid state drive and 16GB of DDR memory. At today's rates, that's USD $187 a piece pre-loaded with Windows 11 and they run great out of the box.
?
For single-threaded stuff like PSUD they are still over 50% of the speed of a Core i9 machine costing 6 times the price. Computing has never been cheaper in real terms :)
?
Regards,
Duncan
?
?
?
?
|
Re: Where is “download PSUD2” now?
On 12/24/2024 8:01 AM, Richard D via groups.io wrote: I don't know much about Appimages or how much work creating them involves, but they seem to run on any OS without installation, It seems to run "on many Linuxes"; no mention of Mac or Win. "It aims to allow the installation of binary software independently of specific Linux distributions... one AppImage can be installed and run across Ubuntu, Arch, and Red Hat<snip> without needing to use different files" And the easier it is for the user, I'd bet the harder it is for the programmer. There used to be a free/not-free tool to run some Windows on Mac and Linux? Wine? Crossover? Parallels? There are virtual machines to run Intel code on other hardware and O/Ses- I can even run Win10 on my Win7 machine. My Oracle VM VirtualBox is getting old and wobbly but there is a newer release. Access to Windows installation ISOs is pretty easy. I am seeing minimal-spec Windows machines under $300, which is not pocket change, but may be less than a serious power supply you want to model. Your neighbors are throwing-out Win-machines: our thrift store has a pile of Win10 Dells the local warehouse upgraded from. Remember that PSDU2 always ran fine on old Win7, which your friends should not be using on today's internet (copy the install file on an uptodate machine to a thumb drive, then sneaker-net to the Win7 box).
|
Re: Where is “download PSUD2” now?
Thanks Duncan,
I don't know much about Appimages or how much work creating them
involves, but they seem to run on any OS without installation,
which would be ideal for PSUD.
Richard.
On 24/12/2024 12:08, Duncan Munro via
groups.io wrote:
toggle quoted message
Show quoted text
Hi Richard,
?
Good question - let me look at this. I've got a few days off
over the Xmas break so hopefully should get some time to
investigate.?
?
If it's doable, it's going to be a zip or tgz that you expand
into your home area somewhere and run from there rather than a
fully fledged package (RPM/DEB etc.), but I guess most Linux
aficionados wouldn't mind that sort of thing too much.
?
Regards,
Duncan
?
?
?
?
|
Re: Where is “download PSUD2” now?
Hi Richard,
?
Good question - let me look at this. I've got a few days off over the Xmas break so hopefully should get some time to investigate.?
?
If it's doable, it's going to be a zip or tgz that you expand into your home area somewhere and run from there rather than a fully fledged package (RPM/DEB etc.), but I guess most Linux aficionados wouldn't mind that sort of thing too much.
?
Regards,
Duncan
?
?
?
?
|
Re: Where is “download PSUD2” now?
Sorry, I should have said " Is there any chance of an Appimage for Linux", but I assume the answer is the same.
?
Regards,
?
Richard.
|
Re: Where is “download PSUD2” now?
Hi,
Is there any chance of an Appimage ?
Richard
On 23/12/2024 21:56, Duncan Munro via
groups.io wrote:
toggle quoted message
Show quoted text
Hi Francois,
?
The MacOS version is no longer available or supported, sorry.
?
Regards,
Duncan
?
|
Re: Where is “download PSUD2” now?
Hi Francois,
?
The MacOS version is no longer available or supported, sorry.
?
Regards,
Duncan
?
|
Where is “download PSUD2” now?
I’m looking to download the software for MAC but somehow I can’t find any software download on? /g/duncanampspsud
|
On 12/6/2024 11:12 AM, Jim VE7RF via groups.io wrote: the Power factor is just The classic definition of Power Factor I learned as a boy applies to Linear components (L R C) and continuous currents (no rectifiers). Same for "lead/lag": does not cover the case of rectifiers. The usual reason to think Power Factor is because a classic Electric Utility Power Meter can measure both Real and Imaginary power. When a customer sucks huge gulps of Imaginary energy (induction motors at light load) the line losses are real and significant, and the company bills for that energy. Traditional Residential customers had only resistance loads (lamps, heaters, stoves) or mild inductance (sub-HP motors), and didn't understand the concept, so were only charged for Real power (but at a higher rate to cover the multitude of little losses in residential service). (I hear some new 'smart' meters do count imaginary energy in residential service, progress!?) Your 16KW max load would cost $4/hour here. OK, if it is 24 hours a day (I was in a discussion about the closing of a broadcast radio system) that's $100/day, $3,100/month, or the salary of an asst manager. OTOH if it is a RadarRange which runs 15 minutes a day to cook large meals that's a dollar, and probably insignificant. Any way they meter it. (But taking 15KW peaks on a 150W average billing will get into a Demand Charge.) Yes, as you move from the raw-rectified AC along to the smooth DC your cap values should increase. I don't have a Law for that, but it seems to work that way.
|
On Fri, Dec 6, 2024 at 01:26 PM, Jim VE7RF wrote:
?
Hi Jim,
?
On a simple FWB rectifier with a high? C filter cap....... is the result? ?RMS? I(C1).....? the same as the AC ripple current going into the cap ????
Yes, that's exactly right. Ripple current limit is specified by the manufacturer to limit the effects of heating so RMS is the way to measure this.
?
On another note ( using the same above config)? I noticed that the RMS xfmr current I (T1)? x RMS V(T1)? ?results in a LOT more rms AC power.....? ? vs just the V(L1) x? I (L1).....? like a lot more, like 53.7% more.
Put another way? DC load power / AC power? = .65? ?Am I assuming the Power factor is just .65??
?
?
This isn't power factor (that's a whole different ball game involving an overall phase difference between V and I), it's the ratio between RMS and average power. For the load, RMS and average will be pretty much the same. For the transformer, the conduction angle can be quite short and this inflates the RMS power (short and peaky signals have a higher ratio of RMS to average, that's just the way the mathematics works).
?
Since there appears to be no leading / lagging..... where is the lousy...'apparent power factor'? coming from ???
?
Again, there's no power factor situation here - that's why you're not seeing any lead/lag. The maximum current will peak during maximum voltage; this is where the voltage difference between the transformer and V(C1) is at its highest. Thus causing the current to flow.
?
For the C1-L-C2 config, I noticed that the C2 cap has to be a bare minimum of at least the same value as the C1 cap....and preferably a lot more.? ?If the C2 cap is lower than the C1 cap value, the output DC? voltage oscillates wildly, till it finally settles down
?
That's correct, making C2 bigger than C1 will damp down the oscillations and lower the resonant frequency. C1-L-C2 is effectively a tuned tank circuit, so will always be resonant to a greater or lesser degree.
?
Hope the above is helpful. As a general suggestion, try and make your conduction angle larger - check I(T1) as short high current spikes are unhelpful. Try making C1 smaller still, allow plenty of ripple on C1 as the combo of L/C2 will cut the mains hum down to get the conduction angle up. The larger conduction angle will (tend to) make the power supply more efficient .
?
Regards,
Duncan
?
|
On a simple FWB rectifier with a high? C filter cap....... is the result? ?RMS? I(C1).....? the same as the AC ripple current going into the cap ????
On another note ( using the same above config)? I noticed that the RMS xfmr current I (T1)? x RMS V(T1)? ?results in a LOT more rms AC power.....? ? vs just the V(L1) x? I (L1).....? like a lot more, like 53.7% more.
Put another way? DC load power / AC power? = .65? ?Am I assuming the Power factor is just .65??
On another note ( using the same above? FWB + filter cap config).....when comparing V (T1)? ?vs I (T1)? ...under full load..... that the peak of the current pulse sits dead center in the middle of the (now flattened to a straight line)? peak of the voltage.?
I believe the flattened? Voltage portion is what is causing the harmonics going back into the mains V.?
It also appears that there is no leading / lagging going on.? The top of the current pulse is dead center in the flattened portion of the top of the voltage peak.?
Since there appears to be no leading / lagging..... where is the lousy...'apparent power factor'? coming from ???
The config is as follows.?
Plate xfmr? with 240 vac input? ( 60 hz)....and a 5225 vac secondary ( no load).? Hypersil C core xfmr.?
.1 ohm dc pri resistance.? And a 6.0 ohm dc sec resistance.? ?Source resistance = 53.39 ohms.? Sec winding rated at 3 amps CCS.? ?Xfmr weight = 253 lbs.?
FWB consists of 24 x 6A10 diodes ) 1 kv piv / 6 amps CCS / 400 amps surge rated)? in series? for EACH of the 4 x legs of the FWB.? ? Then two identical FWB assy's are wired in parallel. 192 x diodes in total used.
Filter cap is? 800 uf? ( 25 x 10,000uf? @ 450 vdc lytics in series, each with a 100k @ 3 watt, 1%? MOF across each cap? ( aka...'eq' resistors).? 2 x banks of caps are wired in parallel ( 400+400 = 800 uf).
Each cap is just .0159 ohm for ESR? @ 120 hz.? ?Total for the 50 x lytics =? .19875 ohms.? Each lytic is 3.0" diam x 8.625" tall.. .and weigh 3.3 lbs each. Ripple current rated at 21.6 amps CCS.
I have been using PSUD-2 in various high voltage configurations for eons now.? FWCT, FWB, FWD.? ?Plane C filter....and also a C-L-C? configuration .? ?For the C1-L-C2 config, I noticed that the C2 cap has to be a bare minimum of at least the same value as the C1 cap....and preferably a lot more.? ?If the C2 cap is lower than the C1 cap value, the output DC? voltage oscillates wildly, till it finally settles down.? ?I also have the option of shunting out the L ( via a vacuum relay) to turn it into one big C filter.? ?I also installed an adjustable spark gap across the 4H choke terminals.? ( choke is also a hypersil C core type, 127 lbs, and rated for 3 amps CCS? and 10 kvdc).? ?For safety purposes, the entire choke is well insulated from the bottom of the rack cabinet.?
I have built plenty of HV? B+ supplies over the years.? I'm just not clear on the 'apparent PF' issue, V flattening / harmonics . AC ripple measurements.
Any insight would be greatly appreciated.?
|
Re: What happened to the macos files folder?
Hi Miguel,
?
The MacOS version was withdrawn - it's no longer distributed or supported.
?
Regards,
Duncan
?
|
What happened to the macos files folder?
Hi,?
I've been trying to find the dmg build files but it seems like the folder has been removed from the files section? Duncan can you please reupload it?
?
Best Regards
?
Miguel
|
Re: How to add loads at different points along the filter chain?
Duncan,
This is a very nice feature that I wasn't aware of.
Thanks for making the Power Supply Designer available and for your attentive replies whenever questions arise.
Cheers,
Tom
|
Re: How to add loads at different points along the filter chain?
Hi, you can insert a current tap in various places in the chain. This would allow you to put a tap in for the power amp, and use the final load with a high value resistor (or constant current) to model the draw from the preamp.
?
It might look something like this:
?
?
My example above is shown as LC followed by RC, but you could have 2 x LC instead. Do the filters first, then right click on the lower part of the second filter to get the popup menu. From there you can insert the current tap:
?
?
?
You can then right-click -> Edit on the current tap itself (or double-click) to change the amount of current draw. Hope this helps.
?
Regards,
Duncan
?
|
How to add loads at different points along the filter chain?
I am designing a PSU for a valve amp,? ?and I want to supply the power amp from after the first LC filter stage,? and the preamp from after the next RC filter stage of the power supply.? What is the correct way to add the loads in the simulation schematic?
|
Re: PSUD3 Windows Beta / Preview edition
|
Re: PSUD3 Windows Beta / Preview edition
Hi John,
?
Thanks for the offer; I'll be keeping it as a closed-source project (I do have plenty of open source ones), so wouldn't need any code work.
?
What would help is with some testing of the app when I do a re-release if you're interested in being involved with that?
?
Regards,
Duncan
?
|
Paul Reid