The diyAudio build guides

Only Nelson Pass Amp Kit Available is the Amp Camp Amp (ACA)

http://www.diyaudio.com/forums/pass-labs/247361-aca-illustrated-build-guide.html

Out of stock at the store ATM but should be back in about 3 months:
Amp Camp Amp Complete Kit - Parts, Chassis, 2xPower Supplies - Amp Camp Kit - Kits

Thanks Variac I've seen this, and while I didn't think it wasn't a great deal, I've already built an amp and when I was going to do another, it was going to be an attempt at something 'right up there'. So this amp didn't seem to fit the bill - but it got me to thinking about 'watts' vs speaker efficiency'.

This is only a 5 watt amp. It might spec out pretty good, but if I'm having trouble getting decent volume out of my 85+ watt Virtue M-901 using my 91db speakers, how much of a problem will it be with 5 watts?

But now I'm finishing off a set of 96db open baffle speakers. I have a few other amps including a Elekit TU-879S 6L6GC rated at 8.5 watts that I'm hoping to get better performance out of verses the older speakers.

So the wonder i'm wondering is how does speaker efficiency fit into the amp output curve? How is 5 watts vs 100 watts into 90 db compare to 5 watts vs 100 watts into a speaker even 1 db more efficient?

Dave
 
Working in dB makes the estimation very easy.

1W = 0dBW. This is our reference.
2W = ~+3dBW
4W = ~+6dBW
8W = ~+9dBW
10W = +10dBW
16W = ~+12dBW
20W = ~+13dBW
32W = ~+15dBW
40W = ~+16dBW
64W = ~+18dBW
100W = +20dBW

Note that 1W, 10W and 100W are exact.
Whereas the doubled values are approximate. This is due to double not being exactly 3dB. Double Power is actually ~+3.0103dB

Using speaker sensitivity, add on the dBW of the amplifier and you end up with the SPL output, provided neither the amplifier nor the speaker are overloaded. One also has to make a correction for power compression.

eg
88dB/W @ 1m speaker with a 64W amplifier gives 88 + 18 = 106dB @ 1m

99dB/W @ 1m speaker with a 5W amplifier gives 99 + ~7 = ~106dB @ 1m
 
Last edited:
Thanks Andrew, So most of the designs are different but none of them have any set values that cant be adjusted for more gain?

I was curious because looking at building something like a 30watt amp and then jumping up to something more extreme I wondered if it could be the same basic design or not. Obviously different values for each.

Secondly what is the biggest amp that anyone has created from one of these Pass designs, gain wise?

Hoping that if I were to build a very large output amplifier that perhaps someone already has done this before. If not then I will have to do a LOT more reading.:)
 
I will be building all new speakers for my front HT channels. The center will be around 100db or so and the L/R will be around 98db. So with music listening and movie watching I wont need much power. BUT I have always wanted to build a much larger amplifier for my subs. I will be using pro audio 18's in simple sealed cabinets. LT circuit and thats all. BUT they need a LOT of power to hit Xmax. (18 Sound 18LW2500)

So for the first amp I will only need about 16-30watts and be plenty happy. I could use a simple 4watt amp on the center and be happy also. I am mainly looking at having enough power for music listening. Add in EQ and I will want to have some headroom leftover. So this will be the first type pf build I would like to do. Was looking at your guides and the F4 or possibly F5v2(I believe its called) might be my first try.

The subwoofer amp will have to slowly come together. Gathering parts and reading as much as possible. I dont think I will be able to afford to build an amp that is big enough to punish my 18's but half way would be nice. I dont intend on breaking anything I just like to have 3db's or more of headroom if I can in all my amps. And for now I only have Class D and Class H amps that I have used.

So this is what I am going by. BUT I also might stick with pro audio amps for the subs and these Pass designs for smaller wattage needs.
 
Last edited:
........... The center will be around 100db or so and the L/R will be around 98db. .................
So for the first amp I will only need about 16-30watts and be plenty happy.................
50W to 60W into 98dB/W speakers are likely to give adequately loud reproduction on very dynamic audio.
You might get away with as little as 20W to 30W into the 100dB/W speaker.

These maximum powers will allow transient peaks to get close to 110dB @ 2.5m listening distance.

A single 18" will struggle to get close to these levels unless you keep it very close to you and it has a highish Xmax..
 
Yes I am not so worried about the subwoofer. I know I would have to have a MONSTER of an amp for that one. And having plenty of headroom for dynamics n a movie and Crest factor in music has always had me wanting much more than is probably needed.

BUT when building these amps I dont want to build something that is never going to happen because of parts costs and complexity. While it would be fun to build a XA1000, lol, its never going to happen. I know it wont because of other reasons but that was just my point.

BUT back to your point I would be about 2.5m - 4m away when watching movies. BUT I dont listen at reference level so 30watts may do it. Or I just double that to make sure I do have enough. And I will have a pair of 18's that will have to have a amplifier channel each going to them or mono. Those re about 2.5meters away. And I will be adding another single 21" down the road. BUT I dont want to talk about speakers just getting power to them is all.:D

F5Turbo sounded like an option but it seems like all designs can be ramped up for whatever output.
 
Thats what I was reading. At what point or rail voltage or other does the design have to change and incorporate a lot more things because of the amount of wattage being produced?

Or is the design simple enough that you just have to change value of components to reach a persons wants?
 
For ClassAB amplifiers the output device total Pmax is approximately 4times to 6times the maximum output power. i.e. for 80W maximum output, use one pair of 150W to 250W devices

For ClassA the device temperature is set by the device dissipation and this determines the maximum ClassA output power.
For the F5 and F5t the maximum device dissipation is approximately 30W for long term reliability.
That 30W sets the limit of the product of ClassA bias current and supply rail voltage.
examples:
1A bias and 30Vcc = 30W device dissipation
1.5A & 20Vcc = 30W
2A & 15Vcc = 30W
0.75A & 40Vcc = 30W
0.667A & 45Vcc = 30W
0.5A & 60Vcc = 30W

You need to set a target and then CHOOSE what combination achieves your target.
 
Last edited:
Thats what I was reading. At what point or rail voltage or other does the design have to change and incorporate a lot more things because of the amount of wattage being produced?

Or is the design simple enough that you just have to change value of components to reach a persons wants?

Alot more things are dependent on the size of the toroid, and on the size of the capacitor bank. As far a rail voltage goes. you either cascode or you don't. You must cascode if the rail voltage is higher than Vds of the 2sk170/2sj74.

Your wattage will depend on rail voltage.
 
Yes thanks for that. Havent chatted with 6L6 more on questions I had I now know what I will need for my wants. For me the F5v3 would be best for me. BUT I just need to find the space for it. I will slowly collect parts for one of these builds. And hopefully when its time to solder I will have the room. :D
 
Halfway there! We had some technical difficulties for this build. I didn't see the note about pre-setting the pots until after they were soldered to the boards and then I could not measure them since they were in-circuit.

We powered it up with a Variac to 50V and dialed out both channel offsets to 0V and the bias to 0.35V drop. I then raised it to 70 and one channel behaved fine but the other my bias went to the negative rail (-21V). My fear is that R8 was at one extreme end of its range ... would this fry the JFETs?

I ended up removing the R8 pot from both channels and replacing with a 1K and the good channel still functioned fine but I still get -21V offset on the bad channel and adjusting R7 does nothing. I can adjust bias but not offset. I also noticed the JFETs are VERY hot.

I removed the V+ and V- from the bad channel, set the offset and bias on the good channel and listened and it sounds good albeit veiled. Of course it needs to break-in, but so far so good.

Any ideas on the bad channel? Did I toast the JFETs?
 

Attachments

  • halfway.JPG
    halfway.JPG
    430.6 KB · Views: 1,519