You can use mjl3281/1302 , mjl 21193/4,njl0381/0302 (mine)
or 2sc/a 5200/1943..pretty much any NPN/PNP pair out there
(to-264 /to-3P package). Drivers can be MJL-xx or 2sa/c 1837/
4793, fast expensive toshiba's..Full choices will be in BOM..
If you are strapped for cash$$, just use 2 op's per amp.
On 1/1/2009 FA1 and 3 will be fully doc'ed with
complete BOM, board,Schema w/readings and silkscreen..
I'll let 2 go, so I can fully devote to 1 and 3..
OS
or 2sc/a 5200/1943..pretty much any NPN/PNP pair out there
(to-264 /to-3P package). Drivers can be MJL-xx or 2sa/c 1837/
4793, fast expensive toshiba's..Full choices will be in BOM..
If you are strapped for cash$$, just use 2 op's per amp.
On 1/1/2009 FA1 and 3 will be fully doc'ed with
complete BOM, board,Schema w/readings and silkscreen..
I'll let 2 go, so I can fully devote to 1 and 3..
OS
ostripper said:You can use mjl3281/1302 , mjl 21193/4,njl0381/0302 (mine)
or 2sc/a 5200/1943..pretty much any NPN/PNP pair out there
(to-264 /to-3P package). Drivers can be MJL-xx or 2sa/c 1837/
4793, fast expensive toshiba's..Full choices will be in BOM..
If you are strapped for cash$$, just use 2 op's per amp.
On 1/1/2009 FA1 and 3 will be fully doc'ed with
complete BOM, board,Schema w/readings and silkscreen..
I'll let 2 go, so I can fully devote to 1 and 3..
OS
Information that I've collected so far:
Widlar
Bootstrap
51vdc rails (sweet price point)
From this post (#2 on this thread) http://www.diyaudio.com/forums/showthread.php?postid=1693313#post1693313
Parts that have already arrived in the mail (hoping these aren't a mistake):
NJW3281G, NJW1302G http://www.onsemi.com/pub_link/Collateral/NJW3281-D.PDF
MJE15032, MJE15033 http://www.onsemi.com/pub_link/Collateral/MJE15032-D.PDF
BD139, BD140
So, do I need to return all of these in order to match the new schematic?
keantoken said:The simulator will teach you a LOT about discrete design of you're clever enough to consider all possibilities for a given output. With an open mind, a lot can be gained. Beware, however, don't go putting 300W through a 2N3904.
I have very little real-world experience, as most of my experimenting has been done on the simulator. I've never had the money to do anything. In this case DIYAudio has provided a way for me to reality check my circuits, and to make sure I'm being practical while designing them. But on the side I have also received some criticism for this. Whatever you do, try not to annoy people TOO much.
I would say go for it, as working on the simulator can be a very worthwhile experience.
I would recommend LTSpice, from Linear Technologies. It's free, in the Design support section. There is also an active Yahoo Egroup for it where you can ask questions.
You can also write the output of your simulated circuit onto a .wav file for listening.Go to the help file and search for ".wav".
I can help you understand the sim, if you are interested in my assistance I can Email you so you'll have my Email address.
- keantoken
Oh yes! Thank you for the kind offer. Please do Email.
look at schema on "frugal amp thread"
those might work, but , like MJL, I use ksa1381/ksc3505 (300v - fairchild) and fast Ksc 1845/ksa992 (120v - fairchild -small signal)
the BD's might work For VAS but they have less gain and Vceo of
120...close call for +-50V PS .
You could use BD's for CCS/Vbias..perfect!!
Your OP devices are cool.. NJW3281G, NJW1302G
MJE15032, MJE15033 (250 V Vceo)
OS
BD139, BD140
those might work, but , like MJL, I use ksa1381/ksc3505 (300v - fairchild) and fast Ksc 1845/ksa992 (120v - fairchild -small signal)
the BD's might work For VAS but they have less gain and Vceo of
120...close call for +-50V PS .
You could use BD's for CCS/Vbias..perfect!!
Your OP devices are cool.. NJW3281G, NJW1302G
MJE15032, MJE15033 (250 V Vceo)
OS
ostripper said:. . . I use ksa1381/ksc3505 (300v - fairchild) and fast Ksc 1845/ksa992 . . .
On these, availability differs slightly:
KSA1381ESTU http://www.fairchildsemi.com/ds/KS/KSA1381.pdf
KSC3503ESTU http://www.fairchildsemi.com/ds/KS/KSC3503.pdf
and
KSC1845,KSA992 are available in gain figures of: 200 or 300 and I don't know which to choose. I'd assume 300?
Thanks for the news on the output devices.
go and read the spec again.danielwritesbac said:KSC1845,KSA992 are available in gain figures of: 200 or 300 and I don't know which to choose. I'd assume 300?
Does the datasheet show a range of hFE? Min, Typ, Max? or a graph of hFE variations with temperature or with Id or with Vce?
Are different hFE ranges available with different codes?
The NEC version is available in three grades P, F & E and a fourth grade U for the NPN only
AndrewT said:go and read the spec again.
Does the datasheet show a range of hFE? Min, Typ, Max? or a graph of hFE variations with temperature or with Id or with Vce?
Are different hFE ranges available with different codes?
The NEC version is available in three grades P, F & E and a fourth grade U for the NPN only
For local availability, and when both match, its either P or F.
ostripper said:Drivers can be MJL-xx or 2sa/c 1837/
4793, fast expensive toshiba's..
?
Attachments
calculate the effect on performance that the range of grades and within grade will have. Decide if you can operate over that whole range. If you need to control the variation to get your required performance then you have to select the grade that achieves it.danielwritesbac said:and when both match, its either P or F.
If you need even tighter parameter control then matching/selection becomes inevitable.
The lowest grade P should be the cheapest, the matched is the most expensive option.
AndrewT said:nope.
I have amps that run @ +-50.5Vdc and produce 110W into 8r0 and/or 200W into 4r0.
FET amps require more supply voltage to achieve these power outputs.
Is this then too close to maximum for my output transistors?
danielwritesbac said:
Is this then too close to maximum for my output transistors?
Those are 250 volt devices. This means they will operate up to 125V rails.
You should do some reading on SOA (safe operating area) and learn a bit more about the spec sheet for the devices you want to use.
http://www.onsemi.com/pub_link/Collateral/NJW3281-D.PDF
http://www.onsemi.com/pub_link/Collateral/NJW3281-D.PDF
Yes, its probable that I need to use 48vdc rails. No worries. The power supply has setting for either 48vdc or 51vdc.
With OS's estimate of 45vdc rails to produce 70W to 8 ohm speakers, I had expected 51vdc rails to produce 160W to 4 ohm speakers. In this case, the thermal junction stays within tolerances.
With AndrewT's estimate of 51vdc rails to produce 200W to 4 ohm speakers, the thermal junction wouldn't always stay within tolerances and thus I'd need to select a lower voltage.
These considerations are quite simple, with the limiting factor being the size of the thermal interface, not the OnSemi document. The problem was the sudden 40 watt bonus power. That was unexpected.
With OS's estimate of 45vdc rails to produce 70W to 8 ohm speakers, I had expected 51vdc rails to produce 160W to 4 ohm speakers. In this case, the thermal junction stays within tolerances.
With AndrewT's estimate of 51vdc rails to produce 200W to 4 ohm speakers, the thermal junction wouldn't always stay within tolerances and thus I'd need to select a lower voltage.
These considerations are quite simple, with the limiting factor being the size of the thermal interface, not the OnSemi document. The problem was the sudden 40 watt bonus power. That was unexpected.
danielwritesbac said:
These considerations are quite simple, with the limiting factor being the size of the thermal interface, not the OnSemi document. The problem was the sudden 40 watt bonus power. That was unexpected.
The document will tell you all you need to know. One of the limiting factors for power output of an amp is the size of the heatsink. To overcome the dissipation limits of the individual output devices, you can add more, but adding more outputs to an inadequate heatsink will not improve power handling.
Start with a target power output - like 100watts. Source the heatsink that is capable of dissipating the heat produced and then look at the SOA of the devices involved to determine how many output pairs you'll need.
As for the voltage of your power supply, I recommend you leave it high but reduce the gain of the amp to limit the output. This will leave some voltage headroom to prevent clipping during transients.
By DWB - its either P or F.
KSA1381/3503 F is the one. with this grade one does not
need to EF the active side of the VAS (1381) but I might
make the "universal blameless" (FA1)and add pads for EF
in the layout , ability to use a bootstrap or even a
CFP input stage.
OS
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Solid State
- First discrete amp, Need help with NTE 390, 391, 375, 398, and BD140, 139 project