@McQuaide, I'm a relative beginner, but I'll put myself out there anyway. Hopefully, someone will correct me, and I'll learn something.
Taking it step by step:
To make any meaningful assessment, we'd need to know at what voltage the rails will be set to. I'm planning to use the UPSU for an M2X, which is specified to have +/- 25V, so my transformer outputs 18V AC (DC out = AC in * 1.414)
Given these, will the PSU as specified suffice for the F5T V2?
Kind regards,
Drew
Taking it step by step:
To make any meaningful assessment, we'd need to know at what voltage the rails will be set to. I'm planning to use the UPSU for an M2X, which is specified to have +/- 25V, so my transformer outputs 18V AC (DC out = AC in * 1.414)
- 5 x 0R47 resistors: these can be thought of as a buffer betweeen two banks of capacitors. The first bank does the most 'smoothing' , and the second bank goes even further. However, whatever value we place between the banks will reduce the effective power delivered to the amplifier circuit. We don't want to waste a lot of power there, so we put a pretty low value (0R47). Because N resistors in parallel create an effective resistance of R/N, we can reduce the total power loss by adding more resistors. If each resistor can dissipate up to 3W (for example), then each resistor can handle 2.45A (I=sqrt(P/R)). Thus with 5 resistors, the total "safe current" is 5 * 2.45A or about 12A. Adding more resistors will increase the "safe current" limit.
- 22KR bleeder resistor: this resistor is not really for normal operation (i.e. supplying power to the amplifier circuit), but rather to provide a path to "drain" the capacitors when the amp is turned off. A higher value bleeder resistor will decrease the current level, thus increasing the time necessary to discharge the capacitors. Since the current is lower, the power rating can be smaller as well (P = R * I^2). A smaller bleeder will discharge the capacitors faster, but at the cost of higher current, and so they must be rated for a higher power dissipation. Having said that, the bleeder is *always* bleeding, so something higher is desirable so that it doesn't outpace the capacitor banks. You could put two 22KR resistors in parallel to make an effectively 11KR resistance that handles twice the power of a single resistor, and thus drains the capacitors faster.
- 10KR LED resistor: The purpose of this resistor is to limit the current through the LED. Higher current means brighter light, and vice versa. The LED is a diode, so the voltage drop across it is a constant because the supply is DC; in my case it's 1.6V (I think). Given this, we can calculate the current flowing through the LED and the resistor as I = V/R. The voltage across the resistor will be 23.4V (25V - 1.6V), and so the current will be 23.4V/10K = 2.34mA (the LED has very little resistance in this case). Practically speaking, you are controlling the brightness of the LED with the resistor. If it shines too brightly at 10K, increase the resistance to reduce the current. And vice versa.
Given these, will the PSU as specified suffice for the F5T V2?
Kind regards,
Drew
Last edited:
Yes. If you're going to place an order with Mouser, DigiKey, Element14, Arrow anyway -- add one more line item to your order and buy some truly excellent diodes for your F6. They will be cheaper than a ticket to the cinema and they will provide you hundreds more hours of entertainment pleasure.
I assume your F6 will use a 2 x 18 VAC power transformer as recommended in the F6 build guide, and I also assume you will build an unregulated DC power supply having two fullwave rectifier bridges (thus 2 x 4 diodes), as shown in the F6 build guide. Possibly using the diyAudio Store's "Universal Power Supply" PCB, or possibly not.
Sometimes builders forget that each diode in a fullwave bridge circuit, never sees more than (2 x Vout) of reverse voltage. Even during a mains spike; yes really. Since Vout for an F6 is 24 volts, a fullwave bridge rectifier needs four diodes rated for 48 or more volts of reverse potential. For peace of mind, add some safety margin and pick diodes rated >75 volts reverse potential. And sleep soundly.
I prefer diodes with the lowest available forward voltage drop, because those will dissipate the least amount of heat. So they won't cook your nearby electrolytic capacitors and reduce their operating lifetime. Here are some low cost options:
Code:
mfr part num fwd current rev voltage fwd drop USD@qty.1
Littelfuse DST2080C 50 A 80 V 0.60 V $ 1.23
Vishay V40100CI-M3 40 A 100 V 0.64 V $ 1.84
STMicro FERD40H100S 40 A 100 V 0.65 V $ 1.45Attached below is a thermal imager photo of a PSU board with Schottky diodes on heatsinks, mounted fairly close to large electrolytic capacitors. There really is such a thing as diode power dissipation. Also, there really is such a thing as the inverse square law, so move those capacitors as far away as you possibly can.
_
Attachments
Last edited:
I'm building a Honey Badger and have one Universal Power Supply PCB. I'll be using an Antek 45Vx2 800VA transformer.
The Honey Badger build guide suggests a power supply with 40,000uF per rail.
I'm having trouble finding 80V 10,000uF caps that are either in stock or at a decent price. I found some nice 80V 6800uF caps but I'd need to use more to get the similar capacitance. I was thinking 6 of those 6800uF per rail, or 12 total for a dual-rail supply.
Is there an alternate diyaudio PCB, or another recommended PCB, for 6 or more caps per rail in a CRC configuration? I'm trying to simplify my supply with one PS board, one transformer, and one soft-start, if possible.
Any suggestions?
The Honey Badger build guide suggests a power supply with 40,000uF per rail.
I'm having trouble finding 80V 10,000uF caps that are either in stock or at a decent price. I found some nice 80V 6800uF caps but I'd need to use more to get the similar capacitance. I was thinking 6 of those 6800uF per rail, or 12 total for a dual-rail supply.
Is there an alternate diyaudio PCB, or another recommended PCB, for 6 or more caps per rail in a CRC configuration? I'm trying to simplify my supply with one PS board, one transformer, and one soft-start, if possible.
Any suggestions?
Not sure where you/re located or your specific criteria for the caps, but I found at least 10 options <USD $10 per unit of 10kuF (or more) 80V (or more) caps in stock between Mouser and Digikey that would fit into the UPS v3 board you already have.
I know that doesn't answer the question you asked, but it may help.
Once nice example is these;
https://www.mouser.com/ProductDetai...GAEpiMZZMvwFf0viD3Y3T2mGac%2BTKhA/dpqs3b5vEY=
Someone could double check size, lead spacing etc, but I'm fairly certain these would work nicely.
I know that doesn't answer the question you asked, but it may help.
Once nice example is these;
https://www.mouser.com/ProductDetai...GAEpiMZZMvwFf0viD3Y3T2mGac%2BTKhA/dpqs3b5vEY=
Someone could double check size, lead spacing etc, but I'm fairly certain these would work nicely.
Those should work. Thanks! Those didn't show up the last time I searched. Guess I used the wrong search filters.
This answer was really helpful to me! Do you have a finished overhead shot of a power supply that uses both the rectifier section and the capacitor section of the DIY boards? I would like to confirm what the connecting wiring would look like.The "Mains" refer to the wires connected to the AC wall plug. The "Primaries" of the transformer connect to the mains. The transformer then transforms the wall voltage down to out transformer voltage, in a Pass amp something like 18V - 24V. The wires with the lower voltage (which is still AC, just lower voltage) are the "Secondaries"
The lower voltage AC needs to be turned into DC - diode bridges do this - the array of 4 discrete diodes with heatsinks (per channel, totaling 8) take the AC on the secondaries and make it DC, which is not particularly clean and smooth, which is what the filter section is for - The big capacitors and resistor banks will make the rippily DC into very smooth DC.
The "Monolithic bridges" are the little square packages that have 4 diodes in the one package, and take the place of the diode section of the PCB. They have 4 connections so AC goes in, and DC comes out.
In the below photo you can see, from left to right -
Transformer and transformer secondary wires, connected to;
(2) monolithic bridges, connected to;
Filter section of the PCB. (which has the diode section removed)
This page has TONS of awesome info that will help you figure out your PSU - although it's specific to a chipamp, the Pass style PSU is essentially the same, just much bigger. Building a Gainclone chip amp power supply.
Is there any advantage/disadvantage to using monolithinc bridges vs. the DIY diode board?
I am working on this for an F6 but may well want to reuse it in the future for other builds.
Thanks.
Bruce
Here is one example...
https://www.diyaudio.com/community/threads/a-noobs-first-first-watt-m2x.338175/page-2
Edited to add - I can't seem to figure out how to point to one specific post or picture, but there are a few decent pictures of a connected PSU and a PSU under test conditions in that thread. I'm certain there are more somewhere.
Monolithic - Easy, maybe less expensive, generally physically smaller, the First Watt amps used them.
Discreet - Maybe less voltage drop, maybe higher performance, you can play with snubbers on the board, and maybe bragging rights.
Thanks for your response! I guess this is what you were directing me to?
https://www.diyaudio.com/community/attachments/img_1209-jpg.759164/It helps but I still love to see this set up fully wired to transformer, etc. in an F6 specific amp. What do snubbers do?
Any time. Sorry, those are pictures of the various input stages for the M2x amplifier. I'm away from my computer that has the original photos, so I cannot re-attach them.
There are some partial pictures of the PSU starting in #20. There is a complete pic in #22.
You can find a wonderful build guide and pictures for a First Watt PSU in this thread. Note, they built it using monolithic rectifiers. If you're on the fence, and if you're a new builder, I'd recommend monolithic.
https://diyalephj.blogspot.com/
There is a lot of (potentially contentious) discussion around the effectiveness / necessity of snubbers for amplifiers like the F6. So, I will not get into that. What an RC snubber network actually does can be found in the first few sentences of this wonderful thread:
https://www.diyaudio.com/community/...rmer-snubber-using-quasimodo-test-jig.243100/
re: connecting the PSU to your F6 amplifier boards - there should almost definitely be some pictures in the F6 build guide that will help you tremendously. If any of those don't work for you, I'd ask in that thread for help. I've seen a number of color-coded diagrams floating around.
rock12:
At very least, find yourself eight 22,000 uF capacitors; they should be readily available. But consider getting slightly larger caps -- I've found that Nelson's power supply schematics tend to have capacitance than I like. Increasing his suggested capacitance by 50% won't hurt anything (mostly, no bloat) and I suspect you'll appreciate the slightly heftier low end. 27k or 33k uF caps fall in this range. If you're the cautious type, use the 22k uF caps and leave room on the floor of your amp for an additional bank of 2 or 4 caps, which you can add after you've had a chance to grow accustomed to the sound of the amp as originally built.
Naturally, your mileage may vary.
Regards,
Scott
At very least, find yourself eight 22,000 uF capacitors; they should be readily available. But consider getting slightly larger caps -- I've found that Nelson's power supply schematics tend to have capacitance than I like. Increasing his suggested capacitance by 50% won't hurt anything (mostly, no bloat) and I suspect you'll appreciate the slightly heftier low end. 27k or 33k uF caps fall in this range. If you're the cautious type, use the 22k uF caps and leave room on the floor of your amp for an additional bank of 2 or 4 caps, which you can add after you've had a chance to grow accustomed to the sound of the amp as originally built.
Naturally, your mileage may vary.
Regards,
Scott