So... Nico made a observation in the other thread that if your source is AC coupled, the input cap (C1) on the amp may be series with the output cap on the source. And this could have the effect of lowering the net value of the capacitor and therefore impacting low frequency response by raising the corner frequency of the input high pass CR filter.
The board has enough space to expand the input connector from 2 position to 3 position. I'm thinking of adding this, with the 3rd position used to bypass the input cap, making it DC coupled.
The idea being that a builder could add a slide switch to the back panel to allow the connection from the source to be AC or DC coupled. I'd suggest the switch be the shorting variety (make before break), which would allow it to be switched while it's active.
Optional of course. You could build it the conventional way.
Thoughts?
The board has enough space to expand the input connector from 2 position to 3 position. I'm thinking of adding this, with the 3rd position used to bypass the input cap, making it DC coupled.
The idea being that a builder could add a slide switch to the back panel to allow the connection from the source to be AC or DC coupled. I'd suggest the switch be the shorting variety (make before break), which would allow it to be switched while it's active.
Optional of course. You could build it the conventional way.
Thoughts?
Smart move. The user can decide if he wants bass extension. Truly DC coupled amp. Better in everything sonic. Improved phase response, improved distortion. Simply great!
I HOPE YOU CAN STILL GET THIS AMP INTO THE MARKET SUCCESSFULLY, IT NEEDS A GOOD NAME. HOW ABOUT THE BRIAN add Coyote in the logo
I HOPE YOU CAN STILL GET THIS AMP INTO THE MARKET SUCCESSFULLY, IT NEEDS A GOOD NAME. HOW ABOUT THE BRIAN add Coyote in the logo
Last edited:
Fantastic option to add. Several of my ‘vintage’ amps have this option and most times I choose DC coupled. 😉
Haha. I've been referring to it as the "Ultra +P". The "+P" to hint at the extra power that the boosted IPS/VAS rails allow for. Both in terms of Class AB power by swinging closer to the rails and Class A power by reducing dissipation thereby allowing for more bias.
It's borrowed from the firearms industry where "+P" ammunition is for loads with extra power.
After adding the 3 position input, I was looking the layout over and decided to make a tweak to the a few component positions. I changed all the pots to have the same orientation and cleared the components obstructing them from the top/bottom. The idea is to allow for the use of side adjustment pots rather than the top adjust pots shown in the 3D rendering.
This will allow a builder to adjust the DC offset and Bias from the top once the boards are installed in the chassis. The boards will likely be implemented with the inputs to the rear, so one board will face up and the other down. You'd just need to have the pots mounted to account for this, and then maintenance will be easier once its all put together.
This will allow a builder to adjust the DC offset and Bias from the top once the boards are installed in the chassis. The boards will likely be implemented with the inputs to the rear, so one board will face up and the other down. You'd just need to have the pots mounted to account for this, and then maintenance will be easier once its all put together.
C5 should be omitted there is nothing for it to do. I would only have a duty with a bias spreader, but also then much larger say 100uF.
Gentlemen,
I can't admit to giving C5 a lot of thought. 100nF seemed reasonable to me, but I can't say why. I seem to recall reading somewhere that the value for the cap used in this position isn't critical. I choose 100nF to make purchasing parts easier as there's already several 100nF poly caps in the build.
I did tinker with different values in SIMs and didn't see any impact (not that simulations are the ultimate truth). Which is another reason I opted for 100nF.
It didn't seem to have an impact on the as-built design. It performed very close to what the simulations predicted. Based on my rudimentary abilities to measure.
With all that said, it sounds like you are both suggesting increasing the value here. I do see Mooly's Lateral MOSFET amp uses 22uF/25V in this position. That is a proven design with decades of use.
The board has plenty of room in the area. I can switch the footprint to a 6.3mm or 8mm electrolytic. 8mm would allow for up to 470uF / 35V.
Thoughts?
I can't admit to giving C5 a lot of thought. 100nF seemed reasonable to me, but I can't say why. I seem to recall reading somewhere that the value for the cap used in this position isn't critical. I choose 100nF to make purchasing parts easier as there's already several 100nF poly caps in the build.
I did tinker with different values in SIMs and didn't see any impact (not that simulations are the ultimate truth). Which is another reason I opted for 100nF.
It didn't seem to have an impact on the as-built design. It performed very close to what the simulations predicted. Based on my rudimentary abilities to measure.
With all that said, it sounds like you are both suggesting increasing the value here. I do see Mooly's Lateral MOSFET amp uses 22uF/25V in this position. That is a proven design with decades of use.
The board has plenty of room in the area. I can switch the footprint to a 6.3mm or 8mm electrolytic. 8mm would allow for up to 470uF / 35V.
Thoughts?
Brian,
re C5. This capacitor is a good idea as it improves the drive to the gates at high frequencies. Any value between 100nF and 10uF is OK. Any bigger than 10uF will cause problems if the amplifier goes into clipping, as it will discharge and then take some time to charge up again. I use a 1uF film cap here, but 100nF will be fine.
I notice that the capacitors from the mosfet gates to the drains have been removed. I normally use 33pF NP0 ceramics here. The only time I have destroyed a lateral mosfet is when I didn't fit these (though I suspect the reason was due to clumsy fingers).
Some time ago when profusion made both Exicon and Alfet laterals, Alfet had a design guide that recommended gate resistors of 300 ohm for the N channel, and 200 ohm for the P channel. This is what I use, and all my amps seem to work fairly well (I have build 4 different lateral mosfet amps over the last few years). For the a resistor I use a 1206 thin film resistor. If you are using through hole, I suggest a 250mW thin film. If the mosfet happens to get destroyed and tries to pass lots of current through the gate (unlikely I know), you want this resistor the act like a fuse and protect the rest of the amp (hence 250mW).
If you plan to run 600mA bias current, then I'd use dual die mosfets. When buying the Exicon mosfets, I suggest you pay the extra and get the selected ones, as this means both channels will bias up the same. Also buy a pair for testing, and a spare pair to put away in case you ever need to replace a pair.
Also I'd add a 330pF film cap from `IN+' to ground to act as an RF filter. A 1 megohm resistor between the input pins J1/1 and J1/2 would be a good idea as it will ensure that C1 is discharged when the input cables are plugged in.
The compensation capacitor C15 (1.5pF) could also be connected to the collector of Q5 (or Q6) rather than the output (you'll need to redo all your loop gain plots). When C15 is connected to the output, it provides a path for Rf that is picked up by the speaker cable to be feed into the input stage. Connecting it to the drivers stops this.
I like the way you have arranged the two mosfets on the PCB.
Paul Bysouth
re C5. This capacitor is a good idea as it improves the drive to the gates at high frequencies. Any value between 100nF and 10uF is OK. Any bigger than 10uF will cause problems if the amplifier goes into clipping, as it will discharge and then take some time to charge up again. I use a 1uF film cap here, but 100nF will be fine.
I notice that the capacitors from the mosfet gates to the drains have been removed. I normally use 33pF NP0 ceramics here. The only time I have destroyed a lateral mosfet is when I didn't fit these (though I suspect the reason was due to clumsy fingers).
Some time ago when profusion made both Exicon and Alfet laterals, Alfet had a design guide that recommended gate resistors of 300 ohm for the N channel, and 200 ohm for the P channel. This is what I use, and all my amps seem to work fairly well (I have build 4 different lateral mosfet amps over the last few years). For the a resistor I use a 1206 thin film resistor. If you are using through hole, I suggest a 250mW thin film. If the mosfet happens to get destroyed and tries to pass lots of current through the gate (unlikely I know), you want this resistor the act like a fuse and protect the rest of the amp (hence 250mW).
If you plan to run 600mA bias current, then I'd use dual die mosfets. When buying the Exicon mosfets, I suggest you pay the extra and get the selected ones, as this means both channels will bias up the same. Also buy a pair for testing, and a spare pair to put away in case you ever need to replace a pair.
Also I'd add a 330pF film cap from `IN+' to ground to act as an RF filter. A 1 megohm resistor between the input pins J1/1 and J1/2 would be a good idea as it will ensure that C1 is discharged when the input cables are plugged in.
The compensation capacitor C15 (1.5pF) could also be connected to the collector of Q5 (or Q6) rather than the output (you'll need to redo all your loop gain plots). When C15 is connected to the output, it provides a path for Rf that is picked up by the speaker cable to be feed into the input stage. Connecting it to the drivers stops this.
I like the way you have arranged the two mosfets on the PCB.
Paul Bysouth
Brain I would ignore above comments as there are no technical evidence for his claims. There may be some consideration of adding a 100pF cap across the input only valid for countries still uses AM radio transmission in >100kHz range. Regarding 600 mA bias and two devices is BS obviously he does not know of SOA. Late comers always seek attention drawn to them.
Your design was bread boarded on Saturday even omitting parts that I did not believe was needed and there was no issues to be found. I am starting with my PCB which would be two amps (one mirrored) following your PCB as left channel on a single PCB as mentioned before. I already have a box made up for it from a project I did not complete. I am using a fixed bias by replacing the bias pot with a 1N4004 diode.
Attachments
Last edited:
I believe adding a 100R resistor in series with C15 would be another option to guard against this.
I think the revised board is ready for production. I made a few more updates to the connectors.
- I reduced the footprint for C1 from 37.5mm to 27.5mm. This allows the input connector to be turned 90 degrees. This should make it easier to route wiring in a 2U chassis.
- I also moved the 9 pin protection connector to the edge of the board to make wiring cleaner.
- As previously mentioned, the trimmers are all vertical mounted now without obstructions to allow side adjustment trimmers. This allows for easier adjustments while mounted in the chassis.
Brian, Mortem
Job well done. I am restarting my Dual Stereo board after seeing this. I take it that you did not use autorouter.
It is not that I don't trust them I just love the punishment to manually route and with the last track you have no idea where to put it and you start over.
EDIT: Maybe, since you have done all the work, I will just use your board and flip one so that my inputs are at the left and right edges so I can use very short connections to the input connectors.
My back plate will be solid 8mm aluminum sheet with two smaller heat sinks on the outside bolted on it so I have space on either side and center for my connectors.
Job well done. I am restarting my Dual Stereo board after seeing this. I take it that you did not use autorouter.
It is not that I don't trust them I just love the punishment to manually route and with the last track you have no idea where to put it and you start over.
EDIT: Maybe, since you have done all the work, I will just use your board and flip one so that my inputs are at the left and right edges so I can use very short connections to the input connectors.
My back plate will be solid 8mm aluminum sheet with two smaller heat sinks on the outside bolted on it so I have space on either side and center for my connectors.
Last edited: