Hi Kle,
it looks like you have dual rectifiers into parallel caps to a single +-PSU feeding the 2 channel GB300.
Since you already have two rectifiers, can I suggest you try two sets of parallel caps. One set of four dedicated to each channel of the GB300. There appears to be sufficient room to stack the caps in the middle and keep the PSU very compact.
I believe you may find an improvement in the performance due to less interaction between channels.
SKA may benefit less than some other power amps due to the designer's claims for very high PSRR, but it is still worth trying for the cost of four caps.
BTW. what size are your caps?
it looks like you have dual rectifiers into parallel caps to a single +-PSU feeding the 2 channel GB300.
Since you already have two rectifiers, can I suggest you try two sets of parallel caps. One set of four dedicated to each channel of the GB300. There appears to be sufficient room to stack the caps in the middle and keep the PSU very compact.
I believe you may find an improvement in the performance due to less interaction between channels.
SKA may benefit less than some other power amps due to the designer's claims for very high PSRR, but it is still worth trying for the cost of four caps.
BTW. what size are your caps?
Hi japjag
ta, should be soon.
I am running the GB300D's in AB mode with 63V DC rails. According to Greg, that equates to about 200W/8ohms and about 400W/4ohms. 😎
You should really be doing this amp. It is a nice amp.
I am not sure what EVA is on about, wrsto to High Freqs, because I have not heard any problems. Perhaps, she could be more concise with what she is saying, rather than suggesting inuendo?
rgds
ta, should be soon.
I am running the GB300D's in AB mode with 63V DC rails. According to Greg, that equates to about 200W/8ohms and about 400W/4ohms. 😎
You should really be doing this amp. It is a nice amp.
I am not sure what EVA is on about, wrsto to High Freqs, because I have not heard any problems. Perhaps, she could be more concise with what she is saying, rather than suggesting inuendo?
rgds
Hi AndrewT
Actually, the bridges are in parallel and there are two sets of 2 x 8K caps in parallel. By putting the bridges in parallel I can float the ground, which I like doing. If you notice, I have also snubbed the bridges with 22nF PP caps, to significantly reduce zener noise. I have also used 100nF PP caps after the PS caps to help reduce noise artifacts.
Yes, you are right. It would be better to have a more capacitance, but, I built it quickly (within 3 days), with Gregs help, which was muchly appreciated.
rgds
Actually, the bridges are in parallel and there are two sets of 2 x 8K caps in parallel. By putting the bridges in parallel I can float the ground, which I like doing. If you notice, I have also snubbed the bridges with 22nF PP caps, to significantly reduce zener noise. I have also used 100nF PP caps after the PS caps to help reduce noise artifacts.
Yes, you are right. It would be better to have a more capacitance, but, I built it quickly (within 3 days), with Gregs help, which was muchly appreciated.
rgds
Kle: for me its just a matter of when😀 I'm still weighing in if i should do Class A or AB 
I got a two way bookshelf and a three-way floorstander with ribbon tweeters that need some amps and i hope this amp fits the bill. Better get some nice heatsinks though first.
I got a two way bookshelf and a three-way floorstander with ribbon tweeters that need some amps and i hope this amp fits the bill. Better get some nice heatsinks though first.Hi,
what is the impedance of the ribbons?
You may need a very special low impedance capable amp to drive these if they do not have a step down transformer in front of them.
what is the impedance of the ribbons?
You may need a very special low impedance capable amp to drive these if they do not have a step down transformer in front of them.
japjag said:
Hi japjag
Greg's GB300D amp keeps getting better and better, as, they run in. I will be posting a review over the weekend. 😀
They have plenty of power, but, they are current limited, so, as long as the load impedance of your speakers is 3ohms(preferably 4ohms) or higher, there will be no problems.
japjag said:
Hi japjag
I forgot to say nice speakers. 😎
Just a little appertiser ... The GB300D's have a valve like character with ss bass 😀 That's all for the moment, more to come
Ska
kle: thats my only hesitation why i'm not pulling the trigger yet, checking the website though is see: Impedance: 4ohms(minimum 3.2ohms). Whats weird is on the back of my speaker it says 8 ohms
kle: thats my only hesitation why i'm not pulling the trigger yet, checking the website though is see: Impedance: 4ohms(minimum 3.2ohms). Whats weird is on the back of my speaker it says 8 ohms
Re: Ska
If they are 8ohms, then, there does not appear to a problem
It would appear that all is ok 
Hi japjagjapjag said:kle: thats my only hesitation why i'm not pulling the trigger yet, checking the website though is see: Impedance: 4ohms(minimum 3.2ohms). Whats weird is on the back of my speaker it says 8 ohms
If they are 8ohms, then, there does not appear to a problem
It would appear that all is ok well your post regarding the minimum impedance the SKA can handle got me thinking, and i'm aware of impedance dips that happen for speakers which have a claimed 8 ohm impedance so its better to be safe. Well I do hope you post your detailed review...what speakers are you using by the way?
Hi Kle,
you have confirmed you have +-16mF shared between 2channels of your stereo GB300.
You are massively under resourced.
For 8ohms speakers I, and many others, recommend 2mF per amp of peak output current.
If you have an output of 48Vpk then Ipk = 48/8 = 6A . But most 8ohm speakers have a minimum impedance of about 6r so you may consider your Ipk to be as much as 8A.
Now apply the 2mF rule; you need +-12mF to +-16mF per channel to achieve a balanced performance.
For 4ohm speakers all your currents are doubled and all the smoothing caps need to be doubled.
Some constructors say you can/should use 3mF smoothing per amp peak into your minimum impedance and I tend towards this recommendation.
If I were building a stereo GB300 then I would use 2 times +-40mF, about five times what you have.
Do not bother posting your results until after you compare the "before and after" fitting the correct smoothing.
you have confirmed you have +-16mF shared between 2channels of your stereo GB300.
You are massively under resourced.
For 8ohms speakers I, and many others, recommend 2mF per amp of peak output current.
If you have an output of 48Vpk then Ipk = 48/8 = 6A . But most 8ohm speakers have a minimum impedance of about 6r so you may consider your Ipk to be as much as 8A.
Now apply the 2mF rule; you need +-12mF to +-16mF per channel to achieve a balanced performance.
For 4ohm speakers all your currents are doubled and all the smoothing caps need to be doubled.
Some constructors say you can/should use 3mF smoothing per amp peak into your minimum impedance and I tend towards this recommendation.
If I were building a stereo GB300 then I would use 2 times +-40mF, about five times what you have.
Do not bother posting your results until after you compare the "before and after" fitting the correct smoothing.
Hi Japjaq,
you have not confirmed the impedance of the ribbons yet.
They commonly have a 1r to 2r impedance and this is transformed up to 4ohm or 8ohm by some manufacturers.
You need this information before selecting your amplifiers.
Or else, rely on a good advisor/salesman for compatible equipment.
you have not confirmed the impedance of the ribbons yet.
They commonly have a 1r to 2r impedance and this is transformed up to 4ohm or 8ohm by some manufacturers.
You need this information before selecting your amplifiers.
Or else, rely on a good advisor/salesman for compatible equipment.
Power supply capacitance
I am planning 12 x 10KuF total. I think this is 120mF which would be 30mF per rail per channel.
I have heard of other rules by high power pro designers that are much less though. One rule is 1mF per 10W. This works out only to about 12.5mF per rail per channel for a 250W amp.
So around 20mF per rail per channel or less could be reasonable.
Domestic use is very undemanding as high power transients are brief. One is NOT using the amp as an RMS heater!
Capacity required is in joules which involves duration, so Amps can be misleading unless listening to a continuous sine wave!!
cheers
I am planning 12 x 10KuF total. I think this is 120mF which would be 30mF per rail per channel.
I have heard of other rules by high power pro designers that are much less though. One rule is 1mF per 10W. This works out only to about 12.5mF per rail per channel for a 250W amp.
So around 20mF per rail per channel or less could be reasonable.
Domestic use is very undemanding as high power transients are brief. One is NOT using the amp as an RMS heater!
Capacity required is in joules which involves duration, so Amps can be misleading unless listening to a continuous sine wave!!
cheers
HF,
people seem to forget that capacity numbers do not mean much without the voltage rating. A capacitor with twice as much voltage on it has 4 times as much energy stored in it.
imo, the linear rule only applies to class A and layouts for lower impedance use because of ripple levels.
Like you said, a general AB amplifier is not intended for PA abuse.
A linear relationship between PS capacitance and PS voltage would be more logical.
people seem to forget that capacity numbers do not mean much without the voltage rating. A capacitor with twice as much voltage on it has 4 times as much energy stored in it.
imo, the linear rule only applies to class A and layouts for lower impedance use because of ripple levels.
Like you said, a general AB amplifier is not intended for PA abuse.
A linear relationship between PS capacitance and PS voltage would be more logical.
Hi,
Hifryer,
I understand your logic, but I do not agree.
Jacco,
I do not understand the point you are making.
The voltage ripple and/or droop on the power rails is directly proportional to the current delivered to the load.
If you want acceptable ripple then you need adequate smoothing caps. Our disagreement is over adequate and how to define it.
As the peak load current increases, my contention is that the caps should increase proportionately.
It follows from this that halving the load impedance requires double the capacitance. This happens roughly, but only by coincidence, to apply to your mF per watt AT A FIXED SUPPLY VOLTAGE.
As you raise the Vrail voltage a fixed load impedance can draw more current from the supply. Again an increase in load current demands an increase in capacitance IF you want the same level of ripple.
Two noticeable improvements follow from using adequate smoothing:- improved bass response and improved immunity to ripple induced artefacts further up the frequency spectrum. There may be others (increased power is one but the amount is insignificant). Overall a cleaner more accurate sound.
AKSA promotes strict adherance to the philosophy of providing JUST enough of his specification caps for his amps and no more. He claims that excess smoothing promotes poor sound quality. Now go and look at his recommendation and you'll find it is very close to 2mF per Ipk, although he specifies the value somewhat differently.
Hifryer,
I understand your logic, but I do not agree.
Jacco,
I do not understand the point you are making.
The voltage ripple and/or droop on the power rails is directly proportional to the current delivered to the load.
If you want acceptable ripple then you need adequate smoothing caps. Our disagreement is over adequate and how to define it.
As the peak load current increases, my contention is that the caps should increase proportionately.
It follows from this that halving the load impedance requires double the capacitance. This happens roughly, but only by coincidence, to apply to your mF per watt AT A FIXED SUPPLY VOLTAGE.
As you raise the Vrail voltage a fixed load impedance can draw more current from the supply. Again an increase in load current demands an increase in capacitance IF you want the same level of ripple.
Two noticeable improvements follow from using adequate smoothing:- improved bass response and improved immunity to ripple induced artefacts further up the frequency spectrum. There may be others (increased power is one but the amount is insignificant). Overall a cleaner more accurate sound.
AKSA promotes strict adherance to the philosophy of providing JUST enough of his specification caps for his amps and no more. He claims that excess smoothing promotes poor sound quality. Now go and look at his recommendation and you'll find it is very close to 2mF per Ipk, although he specifies the value somewhat differently.
jacco vermeulen said:
Yes , but you are forgetting , that the supply capacitors must be a short for AC signals., and a large capacitor will have a much lower AC impedance at audio frequencies .
AndrewT said:
Hi Andrew
I'm using the BMM CRC supplies with roughly 42mF per channel using the 150w SKA's. Is this enough? According to your estimations its slightly over if I did the calculations right.
The supplies are these one's:
http://www.bmm-electronics.com/Product.asp?Product_ID=3310
For the 300w SKA I'm using these:
http://www.bmm-electronics.com/Product.asp?Product_ID=3311
Which are 82mF per channel
Tube_Dude said:
imo, smaller capacitors in parallel are a better choice for class AB amplifiers.
- Status
- Not open for further replies.