The last period I played a lot with output stages. The last one which entertained me a lot was the Warbler which is a more advanced Mofo.
Initially the Warbler was used with 3.4r/88db bookshelf speakers and needed a lot of bias to give out something.
In the meantime the story with the speakers changed and I got something else that has a minimum impedance of 6r and and a lot more spl, 96db.
Now with these I started to really enjoy the Warbler but unfortunately(for me) another used wanted the modules so had to sell them.
So what to do now?
I was simulating since some time a Warbler with a front end and global feedback so it was time to put it on a pcb.
So this one doesn`t need a preamp in front of it.
Here you can see the modules assembled. The ones from the right are A2Cmx-V power supply filters and negative rail generators.
The amp requires a positive high current power supply and a negative low current one so the A2Cmx-V takes anywhere between 16-32Vdc on its input and gives ~3V less on the high current output and -30V on its negative output.
Here you can see the noise on the output of the amp when powered with laptop brick smps and A2Cmx-V
Initially I went for too much open loop gain and feedback that gave some ringing
The problem was solved by increasing R13
So now was time for some thd testing to see what I get.
With 2A bias you can get a bit more than 10W/8r at quite low thd
For 10wrms a 20v power supply is enough though a bigger power supply voltage will give better thd performance so I opted to do the testing at 29.5V
If you need more power then no problem, you can increase the bias to 3A which will get you 29W/8r which with the right speakers(like in my case) can bring down your roof.
If you don`t dissipate more than 40W you can use to247 devices instead the so227 like I used.
Here is a pic of the amp at work.
The inductor in the picture has 150mH/0.7dcr/4A but actually does 3.6A before starting to saturate. These were quite cheap when compared to hammond, I think half the price.
Depending on the needed power, a 2a/70mH choke is the minimum that can be used with this project, or at least this is what I seen when simulating. I have only these 150mH inductors for test.
If you listened to the Mofo you know already how nice and warm sounding it is, this amp comes close to it but at the same time it has much better control of the speaker.
Initially the Warbler was used with 3.4r/88db bookshelf speakers and needed a lot of bias to give out something.
In the meantime the story with the speakers changed and I got something else that has a minimum impedance of 6r and and a lot more spl, 96db.
Now with these I started to really enjoy the Warbler but unfortunately(for me) another used wanted the modules so had to sell them.
So what to do now?
I was simulating since some time a Warbler with a front end and global feedback so it was time to put it on a pcb.
So this one doesn`t need a preamp in front of it.
Here you can see the modules assembled. The ones from the right are A2Cmx-V power supply filters and negative rail generators.
The amp requires a positive high current power supply and a negative low current one so the A2Cmx-V takes anywhere between 16-32Vdc on its input and gives ~3V less on the high current output and -30V on its negative output.
Here you can see the noise on the output of the amp when powered with laptop brick smps and A2Cmx-V
Initially I went for too much open loop gain and feedback that gave some ringing
The problem was solved by increasing R13
So now was time for some thd testing to see what I get.
With 2A bias you can get a bit more than 10W/8r at quite low thd
For 10wrms a 20v power supply is enough though a bigger power supply voltage will give better thd performance so I opted to do the testing at 29.5V
If you need more power then no problem, you can increase the bias to 3A which will get you 29W/8r which with the right speakers(like in my case) can bring down your roof.
If you don`t dissipate more than 40W you can use to247 devices instead the so227 like I used.
Here is a pic of the amp at work.
The inductor in the picture has 150mH/0.7dcr/4A but actually does 3.6A before starting to saturate. These were quite cheap when compared to hammond, I think half the price.
Depending on the needed power, a 2a/70mH choke is the minimum that can be used with this project, or at least this is what I seen when simulating. I have only these 150mH inductors for test.
If you listened to the Mofo you know already how nice and warm sounding it is, this amp comes close to it but at the same time it has much better control of the speaker.
Last edited:
Thanks ZM!
Still had a few pucks around and the inductors so had to use them somehow.
Now these days its again like this
Still had a few pucks around and the inductors so had to use them somehow.
Now these days its again like this
Had the SITs sitting on my desk waiting to be used with the Gaincore. So today was the day
With these I get a bit lower thd but also a bit less power.
Here it is at 2A bias
And here it is at 3A bias
Because I like what I hear I will keep it like this for awhile 🙂
With these I get a bit lower thd but also a bit less power.
Here it is at 2A bias
And here it is at 3A bias
Because I like what I hear I will keep it like this for awhile 🙂
Recently Pa sent a few lu1014 which I was planning to try in this amp.
So a big thanks to Pa!
I designed some small cascode pcbs that will go together with the lu1014.
No pin strip on the lu1014 pcb because I had bad experience with them in the past so I decided that I will use wires from now on.
So a big thanks to Pa!
I designed some small cascode pcbs that will go together with the lu1014.
No pin strip on the lu1014 pcb because I had bad experience with them in the past so I decided that I will use wires from now on.
I connected them all to my temperature controlled matching jig.
The measurements were taken at 55c heatsink temp, 2.4vds and 1.6a bias.
On the back sticker I have noted the gm, bias current and vgs.
It seems that Pa sent matched devices 😎
The measurements were taken at 55c heatsink temp, 2.4vds and 1.6a bias.
On the back sticker I have noted the gm, bias current and vgs.
It seems that Pa sent matched devices 😎
Thanks!
I have some Tannoy drivers in transmission line enclosures. 6r/96db
I have some Tannoy drivers in transmission line enclosures. 6r/96db
So I went ahead and connected the Lu Triode cell to the Gaincore.
Because of the extra feedback that the cell provides I had to remove some feedback from the amp so I slowed down a bit the Vas by increasing the degeneration on it and also increased the amp closed loop gain to 10x
Without the above modifications it looked like this
Now it looks like this
So at this point I could proceed with some thd testing.
I used to247 mosfets for the cascode devices and these can`t take much dissipation(as the sot227 mosfets or the tokin sits) so I had to lower a bit the bias and the working voltage. Taking this into consideration together with the higher closed loop gain and the lower open loop gain means the amp distorts a bit more than the earlier versions with the puck mosfet and the sit. If you check the graphs you will see that until 10w it does a bit better than the mosfet version and close to the sit version. So at this point I can say the LU is not bad at all and this motivated me to draw a cascode pcb for the sot227 mosfet so I can explore the design at higher power also, not that I need this extra power but at this point it`s difficult to take the candies from the baby 🙂 Thanks Pa! and ZM too!
Later edit:
I forgot to mention but the Triode cell pcbs came together with other pcbs so more fun ahead
Because of the extra feedback that the cell provides I had to remove some feedback from the amp so I slowed down a bit the Vas by increasing the degeneration on it and also increased the amp closed loop gain to 10x
Without the above modifications it looked like this
Now it looks like this
So at this point I could proceed with some thd testing.
I used to247 mosfets for the cascode devices and these can`t take much dissipation(as the sot227 mosfets or the tokin sits) so I had to lower a bit the bias and the working voltage. Taking this into consideration together with the higher closed loop gain and the lower open loop gain means the amp distorts a bit more than the earlier versions with the puck mosfet and the sit. If you check the graphs you will see that until 10w it does a bit better than the mosfet version and close to the sit version. So at this point I can say the LU is not bad at all and this motivated me to draw a cascode pcb for the sot227 mosfet so I can explore the design at higher power also, not that I need this extra power but at this point it`s difficult to take the candies from the baby 🙂 Thanks Pa! and ZM too!
Later edit:
I forgot to mention but the Triode cell pcbs came together with other pcbs so more fun ahead
Last edited:
Ups, missed out this point. I will give it a try with some degeneration too. With heavy you mean around 1r?
fiddle........
I'm not loosing that much sleep chasing triode LU behavior
I mean, not anymore ........
Look at it in other way - if you got desired THD Spectra from complete and intended construction, why bother ?
Practically, you have two iterations to try and choose one - having super pentode LU with higher OLG thus greater NFB, or forcing triode LU with lower OLG and lesser NFB
targeting same CLG and more/less same THD figure
then take in account THD Spectra in broadest possible/full power range and what your ears are telling and, well, how much feet tapping
I'm not loosing that much sleep chasing triode LU behavior
I mean, not anymore ........
Look at it in other way - if you got desired THD Spectra from complete and intended construction, why bother ?
Practically, you have two iterations to try and choose one - having super pentode LU with higher OLG thus greater NFB, or forcing triode LU with lower OLG and lesser NFB
targeting same CLG and more/less same THD figure
then take in account THD Spectra in broadest possible/full power range and what your ears are telling and, well, how much feet tapping
The initial ideea was to use the circuit with the same component values for mosfets, sits and jfets in the os.
Using the triode config for the lu cascode can enable me to use same values for the components as intended initially thus having same(close) olg/clg. I will give it a try cause I like to fiddle.
After this I will be ready to test the new modular design that has 2 new front ends with voltage feedback.
Then there will be enough stuff to play with.
Went silent yesterday.. it was a pretty busy day..
For now I have populated the main pcb that hosts the OS and the C core which is practically same stuff as before only that now is modular.
Then I assembled another core which is the V core, a voltage feedback front end exactly like the one that Pa did for the Stasis.
For now I have populated the main pcb that hosts the OS and the C core which is practically same stuff as before only that now is modular.
Then I assembled another core which is the V core, a voltage feedback front end exactly like the one that Pa did for the Stasis.
I had the Warbler in the past and liked very much its sound though tested it with mosfets and sits.
The P core and the Gaincore OS when put together are exactly the Warbler.
Now that I had the Lus from Pa it was the time to give them a listen in the Warbler 🙂
The P core and the Gaincore OS when put together are exactly the Warbler.
Now that I had the Lus from Pa it was the time to give them a listen in the Warbler 🙂
This time I waned to look at the shape of the 2nd hd.
So here is the Gaincore with the V Core and LU1014 at 2.84V/8r
Here is the Gaincore with the C Core and LU1014 at 2.84V/8r
So here is the Gaincore with the P Core and LU1014 at 2.3V/8r, only 2.3V because the soundcard I use for this synchronized test goes until there.
The P core got it`s name mainly because of the 2nd hd it gives that is Pure 🙂 It gives the highest thd but in a very natural way. The other 2 gain stages are good too but because of the global feedback give much lower thd and alter a bit the shape of the 2nd hd.
All 3 are very musical and at this point I can`t say I have a favorite.
So here is the Gaincore with the V Core and LU1014 at 2.84V/8r
Here is the Gaincore with the C Core and LU1014 at 2.84V/8r
So here is the Gaincore with the P Core and LU1014 at 2.3V/8r, only 2.3V because the soundcard I use for this synchronized test goes until there.
The P core got it`s name mainly because of the 2nd hd it gives that is Pure 🙂 It gives the highest thd but in a very natural way. The other 2 gain stages are good too but because of the global feedback give much lower thd and alter a bit the shape of the 2nd hd.
All 3 are very musical and at this point I can`t say I have a favorite.
Last edited:
Since I mounted the cascoded lu1014 I left them in.
For my daily use I power the amp with ~17.4vdc as rail voltage taken from notebook power supplies. Actually these give 20+ vdc which I filter with an AmyAlice filter followed by a cap multiplier.
Having it powered with the above setup I get 10w/8r with very good sound and performance with a bias of 2.4A so I stopped here because of the dissipation on the to247 device used to cascode the lu1016 starts to get high.
Will try to get to higher dissipation and output power with a sot227 mosfet.
For the moment I play with 2 of the cores, the C core that amplifies the voltage and has current gnfb and the P core which is an impedance adapter that doesn’t amplify the voltage neither has global feedback.
Both have their own character and depending on the music I listen to I like one or the other.
Tonight out of curiosity I decided to take some measurements of the amp with the C core with a 4ohm load.
To my surprise it did quite nice.
For my daily use I power the amp with ~17.4vdc as rail voltage taken from notebook power supplies. Actually these give 20+ vdc which I filter with an AmyAlice filter followed by a cap multiplier.
Having it powered with the above setup I get 10w/8r with very good sound and performance with a bias of 2.4A so I stopped here because of the dissipation on the to247 device used to cascode the lu1016 starts to get high.
Will try to get to higher dissipation and output power with a sot227 mosfet.
For the moment I play with 2 of the cores, the C core that amplifies the voltage and has current gnfb and the P core which is an impedance adapter that doesn’t amplify the voltage neither has global feedback.
Both have their own character and depending on the music I listen to I like one or the other.
Tonight out of curiosity I decided to take some measurements of the amp with the C core with a 4ohm load.
To my surprise it did quite nice.