I am in the process of designing a 120Wrms/60Wrms Guitar amp for 8ohm/16ohm speaker cabinets. I want to try the Sziklai or Complimentary Feedback Pair while employing OnSemi's ThermalTrak BJTs, the initial schematic is attached, so take a look and let me know if you see anything obvious. I like to use ports instead of wiring lines across the entire workspace, so the schematic may be a little harder to follow. PSU will be 38Vrms @ 4.2Arms. The Re's are Caddock MP915-0.050-1%, they are thick film, I noticed most audio builds use metal film. First shot at PA design/build so I'll update as things happen, which won't be all that quick.
Issues to consider:
1) CC capacitor trade-off slew rate for stability
2) Potentiometer in feedback path-RF or prior to RIN for gain(volume) control?
3) Mute attached to some form of protection circuitry and/or front panel.
4) DC offset protection.
5) Short Circuit Protection.
Also, if anyone used frontpanelexpress.com and have a review, lemme know.
Issues to consider:
1) CC capacitor trade-off slew rate for stability
2) Potentiometer in feedback path-RF or prior to RIN for gain(volume) control?
3) Mute attached to some form of protection circuitry and/or front panel.
4) DC offset protection.
5) Short Circuit Protection.
Also, if anyone used frontpanelexpress.com and have a review, lemme know.
Hi Barry,
I haven't yet looked at your diagram, but noted you plan to use those Caddocks for Re. You may want to reconsider that; they have a relatively low overload margin, much less than ceramic or wirewound resistors.
Interestingly, there used to be a Thermal Track from Sanken that had the Re build-in with the same technology as the Caddocks, and that was discontinued because the internal Re blew even when the transistor didn't
jd
The caddocks have to be heat-sinked with thermal grease. I use them on one of the F5's I built. Here's a link to the data sheet: http://www.caddock.com/Online_catalog/Mrktg_Lit/MP9000_Series.pdf
Can you tell from my frequent posts that it's raining here on Saturday morning?
Can you tell from my frequent posts that it's raining here on Saturday morning?
Jack,
It's not the power dissipation, that's OK. But the Caddocks are only (AFAIR)able to survive to 2*design current as short term overload. A 'regular' Re can easily survive, without value change, a 5* short term current overload.
It may never happen, but I thought I would point it out.
PS How's the rain
jd
Hi,
are you using NJLs as drivers?
6mA through the drivers seems very low.
The CFP is very sensitive to Driver junction temperature and relatively insensitive to output junction temperature.
Most do not bother monitoring output temperature at all.
For a PA amp a heatsink temp switch my be better, but it must not latch or you'll lose power until YOU or your Roadie resets it.
The two diodes monitoring the drivers may not compensate well for driver temp variations.
Some who know how to design a CFP stage insist that the problems in making them stable are much reduced if the driver devices are much faster then the output devices.
I seem to recall >150MHz : 30MHz as an example, but don't rely on me, I can't design a CFP.
are you using NJLs as drivers?
6mA through the drivers seems very low.
The CFP is very sensitive to Driver junction temperature and relatively insensitive to output junction temperature.
Most do not bother monitoring output temperature at all.
For a PA amp a heatsink temp switch my be better, but it must not latch or you'll lose power until YOU or your Roadie resets it.
The two diodes monitoring the drivers may not compensate well for driver temp variations.
Some who know how to design a CFP stage insist that the problems in making them stable are much reduced if the driver devices are much faster then the output devices.
I seem to recall >150MHz : 30MHz as an example, but don't rely on me, I can't design a CFP.
Jan, Jack,
Thanks for the discussion on the Caddocks. I have used them for current sense applications before, and they are super easy to mount on heatsinks, I have several laying around, and the power rating (15W) is way overkill for a 0.05ohm Re resistor, but points well taken
Andrew,
Yes I plan on using the NJLs as the Drivers. The LME49810 can source/sink 60mA from the 'Driver Stage' on the chip. As far as Ft is concerned, I will take note of it and alter the design if necessary after the 1st prototype. I have noticed that not very many, if any, designs I have seen posted on this site included a CFP output stage. I learn Heuristically, so at least I will have done, and thus will have learned.
PS Its been raining here all day as well, but its pleasant.
Thanks for the discussion on the Caddocks. I have used them for current sense applications before, and they are super easy to mount on heatsinks, I have several laying around, and the power rating (15W) is way overkill for a 0.05ohm Re resistor, but points well taken
Andrew,
Yes I plan on using the NJLs as the Drivers. The LME49810 can source/sink 60mA from the 'Driver Stage' on the chip. As far as Ft is concerned, I will take note of it and alter the design if necessary after the 1st prototype. I have noticed that not very many, if any, designs I have seen posted on this site included a CFP output stage. I learn Heuristically, so at least I will have done, and thus will have learned.
PS Its been raining here all day as well, but its pleasant.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.