I'm talking from experience here, ie 70's High-Fidelity. Plus as some loudspeaker designers know you can actually increase bass freq extension by tuning a series cap with the woofer.
Thanks HUGH for the big help...I've already done what you said...but R3 value of 33k must be increased to 100k...as what AndrewT suggest or it is okay to retain the value? What will be the effect?
GUYS,......If I add Output cap with 4uf Poly in parallel with 10,000uf as what others do.. does it have sonic effect?
BTW, is 100V rating for output cap be okay or I must increase it to about 150Volts?
GUYS,......If I add Output cap with 4uf Poly in parallel with 10,000uf as what others do.. does it have sonic effect?
BTW, is 100V rating for output cap be okay or I must increase it to about 150Volts?
The output cap voltage is 1/2 the main supply rating in voltage V and capacity uF . Yes I would recommend at least 4uF film cap with output 2K-4K uF electro.
Thanks Infinia.....but for 10,000uf a 10uf Film will be okay?
I had attached my proposed PSU draft with 2 options indicated...
I want comments please>>>
I had attached my proposed PSU draft with 2 options indicated...
I want comments please>>>
Looks good. Consider snubbers or just caps across diodes for RFI/EMI.
RE DC blocking caps. Yes use the same film caps that will be mirrored in a Hi Quality speaker crossover, matching the mid tweeter series section.
Is this for a stereo amp? what if the DC blocking caps value exceeds the PS rail total.
RE DC blocking caps. Yes use the same film caps that will be mirrored in a Hi Quality speaker crossover, matching the mid tweeter series section.
Is this for a stereo amp? what if the DC blocking caps value exceeds the PS rail total.
Last edited:
Yes I intend to use this for stereo at 8 ohms load initially but if it does not enough i can add more caps to lessen the ripple and possible hum.
.....what if the DC blocking caps value exceeds the PS rail total?
R u talking about the ouput coupling cap?...if It is....well .I might as well use a 150V or higher voltage so that in case of shorted output MOSFET it will handle.
.....what if the DC blocking caps value exceeds the PS rail total?
R u talking about the ouput coupling cap?...if It is....well .I might as well use a 150V or higher voltage so that in case of shorted output MOSFET it will handle.
Jun,
Looks good, agree with Infinia. I would use the UFSR diodes, the D92M.
Make sure your 10uH inductor has around 0.1R DCR. This is helpful since it focusses rectifier charge current in one cap, voice coil earth return current in the other. I'd use 0.8mm enamelled wire, at least 80 turns, two layer, air cored. That would be around 2.5m.
R3, base bias resistor, is normally 33K. It should be left the same (though sonically it does not seem to change the sound at all). The LTP stage current is 3ma, So 1.5mA will flow through each LTP device in operation, and with a typical hfe of 600 listed for 2SC1845 this corresponds to a base bias current of just 2.5uA. A 33K resistor will therefore drop just 82.5mV, in practice likely 70-90mV, and this is insignificant compared to the rail voltage, so that an operating point at 80mV less than half Vcc will still result in symmetrical clipping.
The other issue is compensation. A single rail amp will typically require MORE compensation than a dual rail, so I'd change C4 to somewhere between 47pF and 68pF, best found empirically.
BTW, I think R1 should be 1K since it's a low pass filter.
A 10uF filmcap across the output coupler will indeed improve things at the top end. I would suggest using an electro designed for 100KHz switch mode power supply use. These are very good caps, and sound great with an appreciable DC biasing voltage across them. Rate it to at least 100VW. A very large cap is not so important as you might imagine, since higher ESR at low frequencies does not necessarily attenuate the bass response subjectively. In fact it will 'wetten' up the bass, by reducing damping factor at lower frequencies. I would not bother going over 6800uF, to be frank.
Other than that, it's just fine and dandy and will work well, but should have a start up relay on the speaker just as you have indicated. There are other tricks can be done to this very good circuit, but get it built and going first up before we start futzing around with it.
Cheers,
Hugh
Looks good, agree with Infinia. I would use the UFSR diodes, the D92M.
Make sure your 10uH inductor has around 0.1R DCR. This is helpful since it focusses rectifier charge current in one cap, voice coil earth return current in the other. I'd use 0.8mm enamelled wire, at least 80 turns, two layer, air cored. That would be around 2.5m.
R3, base bias resistor, is normally 33K. It should be left the same (though sonically it does not seem to change the sound at all). The LTP stage current is 3ma, So 1.5mA will flow through each LTP device in operation, and with a typical hfe of 600 listed for 2SC1845 this corresponds to a base bias current of just 2.5uA. A 33K resistor will therefore drop just 82.5mV, in practice likely 70-90mV, and this is insignificant compared to the rail voltage, so that an operating point at 80mV less than half Vcc will still result in symmetrical clipping.
The other issue is compensation. A single rail amp will typically require MORE compensation than a dual rail, so I'd change C4 to somewhere between 47pF and 68pF, best found empirically.
BTW, I think R1 should be 1K since it's a low pass filter.
A 10uF filmcap across the output coupler will indeed improve things at the top end. I would suggest using an electro designed for 100KHz switch mode power supply use. These are very good caps, and sound great with an appreciable DC biasing voltage across them. Rate it to at least 100VW. A very large cap is not so important as you might imagine, since higher ESR at low frequencies does not necessarily attenuate the bass response subjectively. In fact it will 'wetten' up the bass, by reducing damping factor at lower frequencies. I would not bother going over 6800uF, to be frank.
Other than that, it's just fine and dandy and will work well, but should have a start up relay on the speaker just as you have indicated. There are other tricks can be done to this very good circuit, but get it built and going first up before we start futzing around with it.
Cheers,
Hugh
Last edited:
Hi Hugh thanks again.....now I can start to build.....however maybe it will take time since I design and etch my own PCB and doing it only during my spare time. Im kinda busy right now...
BTW, can I use these coils specially the big ones in series? these are ferrite inductor.....else i have to build what u had suggested.
BTW, can I use these coils specially the big ones in series? these are ferrite inductor.....else i have to build what u had suggested.
Attachments
Hi Jun,
I would not use ferrite, I would keep it air cored, and inductance low. Large inductance in his position can have very strange sonic effects, best avoided. What is chiefly of interest is the DCR, since this prevents massive rectifier charge currents reaching the output caps, preventing intermodulation with speaker earth return currents. Wind onto a simple plastic former, easy job.......
Keep on truckin', BTW, if you have specific questions about the Quasi design, PM Con, the designer, who is a pal of mine living in Adelaide, South Australia - great guy, super smart and a born leader.
Cheers,
Hugh
I would not use ferrite, I would keep it air cored, and inductance low. Large inductance in his position can have very strange sonic effects, best avoided. What is chiefly of interest is the DCR, since this prevents massive rectifier charge currents reaching the output caps, preventing intermodulation with speaker earth return currents. Wind onto a simple plastic former, easy job.......
Keep on truckin', BTW, if you have specific questions about the Quasi design, PM Con, the designer, who is a pal of mine living in Adelaide, South Australia - great guy, super smart and a born leader.
Cheers,
Hugh
post24, the PSU.
R1 dissipates 2W (140V & 10k). Why?
It will increase the ripple on the supply.
It will consume power. It will heat the chassis ambient.
If you need fast discharge then use a relay to activate R1 only after power is lost.
If you don't need fast discharge then use either none or R1>=100k.
R1 dissipates 2W (140V & 10k). Why?
It will increase the ripple on the supply.
It will consume power. It will heat the chassis ambient.
If you need fast discharge then use a relay to activate R1 only after power is lost.
If you don't need fast discharge then use either none or R1>=100k.
Ah, yes Andrew,
I was looking at R1 on the amp schemat.
You are right. 10K on the power supply is too small a value, it will dissipate 1.96W as you say. It should be 22K 3W, this will have a dissipation of 890mW, so a 3W rating is fine, and it will serve its primary purpose as a bleeder loading down the circuit with a 6.4mA bias.
Hugh
I was looking at R1 on the amp schemat.
You are right. 10K on the power supply is too small a value, it will dissipate 1.96W as you say. It should be 22K 3W, this will have a dissipation of 890mW, so a 3W rating is fine, and it will serve its primary purpose as a bleeder loading down the circuit with a 6.4mA bias.
Hugh
Yes I overlook it has been corrected... Thanks guys
Keep on truckin', BTW, if you have specific questions about the Quasi design, PM Con, the designer, who is a pal of mine living in Adelaide, South Australia - great guy, super smart and a born leader.
Yes I think he is... I had read all the post on that long thread..actually i had already tested to build the lower NMOS version at +- 25 volts and worked good as per listening. However i want it bigger that's why i want this.
Hugh do you knoe where I can contact him? He's not yet posted in the thread and no email at his QUASI site
Keep on truckin', BTW, if you have specific questions about the Quasi design, PM Con, the designer, who is a pal of mine living in Adelaide, South Australia - great guy, super smart and a born leader.
Yes I think he is... I had read all the post on that long thread..actually i had already tested to build the lower NMOS version at +- 25 volts and worked good as per listening. However i want it bigger that's why i want this.
Hugh do you knoe where I can contact him? He's not yet posted in the thread and no email at his QUASI site
Ah, no, Jun, don't have a number but I will comb the SA telephone directory and get back to you,
Hugh
Hugh
Hi Quasi...
Thanks for plugging in......
Yes i want to build an amp with a single rail of 140vdc and I look at your NMOS 350 will fit for this by just having it to work on a single rail. Hugh and AndrewT give me pointers on how to make it work since your not yet in... Now that you're in, maybe you can give me more tips and comments about the proposed modification or if you have other schematics for this i will appreciate it very much.
I have a post at Power amp under development but maybe it tiring to read all the post since its a long thread but it worth a reading.
Thanks for plugging in......
Yes i want to build an amp with a single rail of 140vdc and I look at your NMOS 350 will fit for this by just having it to work on a single rail. Hugh and AndrewT give me pointers on how to make it work since your not yet in... Now that you're in, maybe you can give me more tips and comments about the proposed modification or if you have other schematics for this i will appreciate it very much.
I have a post at Power amp under development but maybe it tiring to read all the post since its a long thread but it worth a reading.
Hi again......another question...
Do I still need to put the Zobel Inductor with as in original NmoS350 ?
I use long wires for my speakers so I usally put inductor on my amps. Does it necessary for a capacitor coupled output? If it does ,what inductance value can you suggest?
Does anybody have a design of protection relay schematics to suit for this type?
Do I still need to put the Zobel Inductor with as in original NmoS350 ?
I use long wires for my speakers so I usally put inductor on my amps. Does it necessary for a capacitor coupled output? If it does ,what inductance value can you suggest?
Does anybody have a design of protection relay schematics to suit for this type?
Thanks Con,
I'd tried in vain to find your phone number, so being here is great for Jun who has very reasonable questions!!
Greetings....
Hugh
I'd tried in vain to find your phone number, so being here is great for Jun who has very reasonable questions!!
Greetings....
Hugh
Hi Hugh...Yes Quasi already drop in this thread....I hope he will reply soon...maybe preparing some schema....hehehe😀
I prefer to play safe and use the full Thiele Network between the amp and the outside world. A Zobel is possibly not enough to guarantee stability or lack of instability artefacts in the output from most amplifiers.
Read JLH 80W design and how he uses a VR+C in the feedback route to correct the scope trace of the output waveform for best sound. Yes, he says a near square wave output without excessively rounded corners and without overshoot sounds better. JLH was a "sound" designer rather than a "numbers" designer.
Great Designers have found that they can design circuits that can stay stable with many of the Thiele Network components omitted.
As I said I play safe and even add a further C+R across the speaker output terminals to create a Pi type filter.
Read JLH 80W design and how he uses a VR+C in the feedback route to correct the scope trace of the output waveform for best sound. Yes, he says a near square wave output without excessively rounded corners and without overshoot sounds better. JLH was a "sound" designer rather than a "numbers" designer.
Great Designers have found that they can design circuits that can stay stable with many of the Thiele Network components omitted.
As I said I play safe and even add a further C+R across the speaker output terminals to create a Pi type filter.
Chosen carefully, a zobel across the voice coil of the speaker reverses the phase shift of the VC inductance thus making life easier for a global negative feedback amplifier.
Looks to me as though you hit the bullseye here, Andrew!
Hugh
Looks to me as though you hit the bullseye here, Andrew!
Hugh
- Status
- Not open for further replies.