How about stability ? I was building an F5t for a friend and it had some issues I had to give up on the build in the end as I spent months on it. I plan to start a new build now using cascode and the new diyaudio PCB,s and take my time without rushing.
Sent from my Lumia 510 using Tapatalk
Sent from my Lumia 510 using Tapatalk
proper cascode
Hey Rixta,
There is a cascode formula within the BOM from the DIYaudio store. Make sure you get and use that and note you must adjust resistors in concert with rail voltage so you MUST plan a build around rail voltage - build your power supply first! Then I think you'll have no issues with cascode. I'm building a v2 with cascode too
Speaking of firsts and I speak from experience, first thing to do is to put a meter on every part you've ordered and make Sure it is what the BOM asks for. It is all too easy to mistake a decimal point, a K for an R etc. I then lable each bag for its place in the BOM. You must also check your soldering along the journey too under a strong light and magnification.
All the problems I have had in my builds could have been avoided with just a little more caution (the caution I am recommending now) and a question or two. It is less fun to build this slowly but now I build without issue and That is More Fun.
Cheers!
Hey Rixta,
There is a cascode formula within the BOM from the DIYaudio store. Make sure you get and use that and note you must adjust resistors in concert with rail voltage so you MUST plan a build around rail voltage - build your power supply first! Then I think you'll have no issues with cascode. I'm building a v2 with cascode too
Speaking of firsts and I speak from experience, first thing to do is to put a meter on every part you've ordered and make Sure it is what the BOM asks for. It is all too easy to mistake a decimal point, a K for an R etc. I then lable each bag for its place in the BOM. You must also check your soldering along the journey too under a strong light and magnification.
All the problems I have had in my builds could have been avoided with just a little more caution (the caution I am recommending now) and a question or two. It is less fun to build this slowly but now I build without issue and That is More Fun.
Cheers!
Hey there. I've been lurking in the background for the last while listening and learning. I've been building up a dual mono F5 V2 with a ton of small changes...ok, some were a little bigger! I hope it still classifies as an f5. I had a bunch of existing parts I have been collecting over the years and just started trying to use for a class a amp. I came across this site and decided to try an f5 turbo. I have 2 650va toroids which once chuffed through the caps and filter provided 60v. I have two like this but they were not center tapped...so I went old school and worked out my layout so I could deal with 60v and ground for my rails and 30v as a virtual ground.
I should have stated that this is my second electrical project (first was a big tesla coil). Any input and advice is welcome although at this point after many fuses, 2 smallish fires and a couple of rebuilds I think I have arrived at a good place!
I've cascoded the front end to deal with the voltages I'm using and set everything up as strict dual mono as possible. 2 juice inputs from the wall, 2 line filters, 2 toroids, 2 bridge rectifiers etc. I built up a large case with some beastly heatsinks that I machined and anodized from a 6 foot extrusion. Because I had the space, and just to make things more complicated I added a b1 preamp into the unit so I guess I technically have an integrated amp. I have tons of pictures to post if anyone is interested and if this beast fits this particular thread.
Cheers
I should have stated that this is my second electrical project (first was a big tesla coil). Any input and advice is welcome although at this point after many fuses, 2 smallish fires and a couple of rebuilds I think I have arrived at a good place!
I've cascoded the front end to deal with the voltages I'm using and set everything up as strict dual mono as possible. 2 juice inputs from the wall, 2 line filters, 2 toroids, 2 bridge rectifiers etc. I built up a large case with some beastly heatsinks that I machined and anodized from a 6 foot extrusion. Because I had the space, and just to make things more complicated I added a b1 preamp into the unit so I guess I technically have an integrated amp. I have tons of pictures to post if anyone is interested and if this beast fits this particular thread.
Cheers
Last edited:
Hi jdg123,
You mentioned the cascode formula within the BOM from the DIYaudio store which I don't have a problem with but I think people may not be using it correctly. The first part rail voltage/(R26+R28) just gives the current through the resistor divider network that supplies bias current to the cascode transistor, for stock voltage of 32v is 2.16 ma. The second part of the equation is the bias resistor times the current which gives the voltage across the bias resistor, which for stock values is 10.3 v, but I think I see people taking this a voltage as being across the cascode transistor but it is not. And it is the current, not voltage through the resistor network take sets bias since this is a BJT and that sets the voltage across the cascode bjt.
Then the question is, what current for each rail voltage to bias the cascode bjt to absorb how much voltage so the voltage across the front end FET is correct.
Just some thoughts that keep coming to mind every time this is discussed.
Have a good one,
John
You mentioned the cascode formula within the BOM from the DIYaudio store which I don't have a problem with but I think people may not be using it correctly. The first part rail voltage/(R26+R28) just gives the current through the resistor divider network that supplies bias current to the cascode transistor, for stock voltage of 32v is 2.16 ma. The second part of the equation is the bias resistor times the current which gives the voltage across the bias resistor, which for stock values is 10.3 v, but I think I see people taking this a voltage as being across the cascode transistor but it is not. And it is the current, not voltage through the resistor network take sets bias since this is a BJT and that sets the voltage across the cascode bjt.
Then the question is, what current for each rail voltage to bias the cascode bjt to absorb how much voltage so the voltage across the front end FET is correct.
Just some thoughts that keep coming to mind every time this is discussed.
Have a good one,
John
JohnH,
you have this all wrong.
The cascode voltage is the voltage at the emitter of the cascode transistor.
The voltage at the base is ~ 600mV above this cascode voltage.
The current flowing through the base stopper, R29 is tiny. The voltage at the divider is virtually the same as the base voltage.
You can safely ignore the effect of this current when calculating the base voltage as a result of the voltage divider effect of R25 & R27
you have this all wrong.
The cascode voltage is the voltage at the emitter of the cascode transistor.
The voltage at the base is ~ 600mV above this cascode voltage.
The current flowing through the base stopper, R29 is tiny. The voltage at the divider is virtually the same as the base voltage.
You can safely ignore the effect of this current when calculating the base voltage as a result of the voltage divider effect of R25 & R27
Hi Andrew,
I knew I could count on you to address this. While all your statements are true none of them contradict anything I wrote and states things I made no mention of.
"The cascode voltage is the voltage at the emitter of the cascode transistor."
Very True, it is HOW to arrive at that voltage I questioned.
"The voltage at the base is ~ 600mV above this cascode voltage."
True again, so to find the cascode voltage you would take the voltage at the voltage divider and subtract 600mV, Now we are getting somewhere.
Thanks for ramming this information through my thick skull. If I under stand this correctly it will help in using different supply voltages for the F5T project when I get to building it.
John
I knew I could count on you to address this. While all your statements are true none of them contradict anything I wrote and states things I made no mention of.
"The cascode voltage is the voltage at the emitter of the cascode transistor."
Very True, it is HOW to arrive at that voltage I questioned.
"The voltage at the base is ~ 600mV above this cascode voltage."
True again, so to find the cascode voltage you would take the voltage at the voltage divider and subtract 600mV, Now we are getting somewhere.
Thanks for ramming this information through my thick skull.
If I under stand this correctly it will help in using different supply voltages for the F5T project when I get to building it.John
Layout of the amp is planned to have the cap bank at the front of the amp, with the FE boards at the back (At the rcas). The power will be distributed to the FE board via the power rails of the output boards.
Should the FE board power rail connection use a pair of decoupling caps? Say 1000uF?
Thanks.
Should the FE board power rail connection use a pair of decoupling caps? Say 1000uF?
Thanks.
Hi All,
After AudioSan call my heatsinks tiny,, I went back and did some calculations on them, they have a 1.34 C/w/8". The temp is stable, 50C at the sink, at the bias I am currently running but if I want to up the bias to where I would really like it then the sinks at idle would be at 65C with the transistor cases at 67C and the Transistor junction Tj at 71C, high but not too bad until you start cranking some volume, then max output has the sinks at 92C, cases at 95C and Tj at 102C. Within the operating limits but too hot for reliability, and this is all the 25C ambient, higher ambients of course make it worse.
So I started looking for a sink in 4" wide with a higher dissapation factor. I finally found one with a .8 C/w/8". Running the calculations again at idle the sinks would be 50C, trans case 52C and Tj 57C. Running at max the sinks would be 67C, case 70.5C and Tj 77.5C, a more acceptable temp and a full 25C drop over the current sinks.
So $478 will get the new sinks here, already cut to length then I have to do a little machining, drilling and tapping, then out to be andonized black. This SA/1 rebuild is costing me a fortune but at least I can reuse the old parts for the F5T.
Have a good one,
John
After AudioSan call my heatsinks tiny,
, I went back and did some calculations on them, they have a 1.34 C/w/8". The temp is stable, 50C at the sink, at the bias I am currently running but if I want to up the bias to where I would really like it then the sinks at idle would be at 65C with the transistor cases at 67C and the Transistor junction Tj at 71C, high but not too bad until you start cranking some volume, then max output has the sinks at 92C, cases at 95C and Tj at 102C. Within the operating limits but too hot for reliability, and this is all the 25C ambient, higher ambients of course make it worse.So I started looking for a sink in 4" wide with a higher dissapation factor. I finally found one with a .8 C/w/8". Running the calculations again at idle the sinks would be 50C, trans case 52C and Tj 57C. Running at max the sinks would be 67C, case 70.5C and Tj 77.5C, a more acceptable temp and a full 25C drop over the current sinks.
So $478 will get the new sinks here, already cut to length then I have to do a little machining, drilling and tapping, then out to be andonized black. This SA/1 rebuild is costing me a fortune but at least I can reuse the old parts for the F5T.
Have a good one,
John
I don't think the output wattage is making the increase in temperature, but simply time. As Sangram says, highest dissipation is at idle, they will actually cool down into a large load.