Why do you say that Eduard? I have two custom made bifillar winded trafos just on the secondaries for the hiragas output with identical secondaries. This can be done easily...
If i remember well Hiraga used a 300va double c core trafo and i use two 300va EI just for the output stage (i better have used two 400va trafos because of heat..). This is allready more than enough even for an 'overclocked' Hiraga amp. But we know that whatever we say here, money talks, and Vishalk didnt sayed what budget he has yet, what power he wants, and what version from the three of them wants.
Cheers
Sergiu
I'm not trying to convince anyone to make an amp just like mine. I just posted some tested facts so that everyone can build whatever he wants and not go through the same testing trials. When i build this amp there werent many people here that conducted tests or posted listening impressions. I had to test almost everything and took conclusions and notes by my own and then invite a couple of friends at blind tests auditions to confirm or not my conclusions.
Cheers
Sergiu
Cheers
Sergiu
@vihalk
i understand you are trying to build a stereo amp. still here is my amp schema for each mono build:
mains EMI Filter [10A rated]
+
soft start module with 2+ sec delay and dual temperature sensor [rated at 75 Degrees Celsius]
+
toroidal trafo : 400VA, +/-20V secondary + small trafo for speaker protection
+
generic bridge rectifier rated at 30A
+
2X10,000uF caps for each rail [40,000uF total]
+
0.25Ohms , 30W ceramic wire wound resistor
+
5X10,000uF + 2X100,000UF per rail [500,000uF total capacitance after resistor]
+
amp board and kit is from Jim's audio. each output transistor are [PNP/NPN] mounted on their own heat-sinks each weighing around 2.7kgs with temperature sensors mounted very close to transistors. audio input is through RCA plug and shielded microphone wire internally.
+
speaker protection
+
banana binding posts
i get about 28VDC from PSU and use star ground on Copper bus bar over capacitors. amp is very very sensitive to PSU and input noise.
probably in future, might try to 'upgrade' the power supply with a choke filter instead of resistor. but for now, amp is absolutely silent - can't hear any noise.
currently, amp is paired with GR BR stand-mount speakers built with Alpair 10.3M wide-band driver. speaker cabinet build is far from perfect but then that is a material for different thread. for now these speakers sound awesome for most parts. if i pair the amp with any other commercially available speakers they sound fabulous.
i understand you are trying to build a stereo amp. still here is my amp schema for each mono build:
mains EMI Filter [10A rated]
+
soft start module with 2+ sec delay and dual temperature sensor [rated at 75 Degrees Celsius]
+
toroidal trafo : 400VA, +/-20V secondary + small trafo for speaker protection
+
generic bridge rectifier rated at 30A
+
2X10,000uF caps for each rail [40,000uF total]
+
0.25Ohms , 30W ceramic wire wound resistor
+
5X10,000uF + 2X100,000UF per rail [500,000uF total capacitance after resistor]
+
amp board and kit is from Jim's audio. each output transistor are [PNP/NPN] mounted on their own heat-sinks each weighing around 2.7kgs with temperature sensors mounted very close to transistors. audio input is through RCA plug and shielded microphone wire internally.
+
speaker protection
+
banana binding posts
i get about 28VDC from PSU and use star ground on Copper bus bar over capacitors. amp is very very sensitive to PSU and input noise.
probably in future, might try to 'upgrade' the power supply with a choke filter instead of resistor. but for now, amp is absolutely silent - can't hear any noise.
currently, amp is paired with GR BR stand-mount speakers built with Alpair 10.3M wide-band driver. speaker cabinet build is far from perfect but then that is a material for different thread. for now these speakers sound awesome for most parts. if i pair the amp with any other commercially available speakers they sound fabulous.
Last edited:
Hello,
If I remember well the voltage did rise well above 24 volts and after a few seconds the voltage did go down to 24 volts.
If you are using caps rated close to 24 volts I would certainly use a bleeder when using a choke input. It could well be possible you need 30-0-30 volts account to get 24 volts dc output.
Like I said before with an r core you might not need soft start
Greetings, eduard
Hello,
Did read some old French magazines and some emails from Vietnam about using LL2733 for choke input with current drawn being similar to the Hiraga amp.
My friend did use a 500 VA 30 volt AC output and connected the two coils of the LL2733 in series to create a 400mH 3,4 ohm input choke with 1,3 A current the ratio between dc output and ac input was 0,76. So one will get 22,8 volt dc.
My friend will start using this supply for a Nelson Pass design in a few months. Right now we dont know if serial or parallel connection will give the best results.
Using transformer with multiple taps will allow to create the same dc output with serial or parallel connected choke.
Will tell more about the French magazines later.....
Greetings, Eduard
Ps. Hope someone will build Hiraga with choke input. Just need to add some extra secundairy taps
Did read some old French magazines and some emails from Vietnam about using LL2733 for choke input with current drawn being similar to the Hiraga amp.
My friend did use a 500 VA 30 volt AC output and connected the two coils of the LL2733 in series to create a 400mH 3,4 ohm input choke with 1,3 A current the ratio between dc output and ac input was 0,76. So one will get 22,8 volt dc.
My friend will start using this supply for a Nelson Pass design in a few months. Right now we dont know if serial or parallel connection will give the best results.
Using transformer with multiple taps will allow to create the same dc output with serial or parallel connected choke.
Will tell more about the French magazines later.....
Greetings, Eduard
Ps. Hope someone will build Hiraga with choke input. Just need to add some extra secundairy taps
Sorry, but this is not true.
High ESR affect low bass - where the most power is necessary, especially on transients.
High needs capacitors with high Rf.
If the Lundahl LL2733 has 0,85ohm ... is a bit too much (10 times), at least for the price they ask.
Actually, both L and Re are too high. I think a much suited value would be 10-30mH and Re</=0,1ohm.
Last edited:
This is what i am planning on doing? Building a Hiraga with Choke input? From the French PDF file you put on this thread?
So how many volts do i actually i need? I thought we discussed using 18-21-24?
Anyway got a quote back from the Transformer Manufacturer, made my eyes water a little! Please attached file.
Got one more quote from another manufacturer!
look what came in the post today 😉
So how many volts do i actually i need? I thought we discussed using 18-21-24?
Anyway got a quote back from the Transformer Manufacturer, made my eyes water a little! Please attached file.
Got one more quote from another manufacturer!
look what came in the post today 😉
I don't think trafo voltage is so critical to be exactly 2x20 or 2x 22 or 2 x 24Vac.
The problem is to have heatsinks big enough to disipate the power.
You must understand that the bigger the trafo voltage is, the bigger the output power will be. And also, the bigger the dissipated power.
In other words, I would see the problem backwards: - if I have heatsinks nr 1 (let's say "recommended") => I can use 2 x 20V trafo; if I have heatsinks nr 2 (much bigger) than I can use 2x24V trafo. And if I have huge heatsinks I can use even 2x28 or 2x30V trafo.
Consider the maximum accepted temp. on heatsink no more than 70C, in the summer.
LE - if you will use coils for filtering than you must calculate how manny volts will be lost on those coils.
Ulost = 2x Ibias x Re coil
Considering Re= 0,85V and Ib = 1,6A => Ulost = 2,72V.
So, for the LL coils add 3V to the initial trafo voltage.
The problem is to have heatsinks big enough to disipate the power.
You must understand that the bigger the trafo voltage is, the bigger the output power will be. And also, the bigger the dissipated power.
In other words, I would see the problem backwards: - if I have heatsinks nr 1 (let's say "recommended") => I can use 2 x 20V trafo; if I have heatsinks nr 2 (much bigger) than I can use 2x24V trafo. And if I have huge heatsinks I can use even 2x28 or 2x30V trafo.
Consider the maximum accepted temp. on heatsink no more than 70C, in the summer.
LE - if you will use coils for filtering than you must calculate how manny volts will be lost on those coils.
Ulost = 2x Ibias x Re coil
Considering Re= 0,85V and Ib = 1,6A => Ulost = 2,72V.
So, for the LL coils add 3V to the initial trafo voltage.
Last edited:
😛 . You're not the only person to mention the importance of heatsinks.
I will be using some MAMMA size heatsinks!
I will be using some MAMMA size heatsinks!
The law of diminishing returns....in class A it makes a big difference. Vishalk, do you know how much power is used at your prefered listening level?
Like it or not, that was the conclusion of extensive testing. I did not do the tests, but they are well documented.
FWIW, transients are not bass, so low ESR should help them, yes.
If you read French you can look up the documentation. The PSU tests are spread out of several issues of la Revue d'Audiophile.
Hello Pano,
We did read the same French publications where they also used a 2 coil choke. It is different than the way the Lundahls are constructed. The French mention that it is important to have identical loads for both + and - side.
Mister Per Lundahl said that it is importasnt for his chokes too.
Jac Music in Germany did write something about the French choke on his website.
@Vishalc. Yes 24-21-18-0-18-21-24 will be the voltages you need to '' play around '' with a CLC supply if you wanna try choke input where the first part after the rectifier is a choke you will some extra windings like 27-30-33 depending on how you will connect the LL2733 400mH or 100mH.
Using a choke as the first element will slow down the charging of the caps at switching on. BUT big transformers depending on if they are EI, R core or toriodal style will make your light dim for a while when switching on. Big toriods 500VA or more are famous for this. Both my R cores 500VA and 650 VA no problem at all.
In the Technical files of Lundahl you can see that if you use choke input the transformer can deliver more current than used with a cap input. The output current from rectifier with choke input is 95% of the transformer specs with capacitor input it will be 63%. One more good reason to give choke input a try!
Greetings, Eduard
We did read the same French publications where they also used a 2 coil choke. It is different than the way the Lundahls are constructed. The French mention that it is important to have identical loads for both + and - side.
Mister Per Lundahl said that it is importasnt for his chokes too.
Jac Music in Germany did write something about the French choke on his website.
@Vishalc. Yes 24-21-18-0-18-21-24 will be the voltages you need to '' play around '' with a CLC supply if you wanna try choke input where the first part after the rectifier is a choke you will some extra windings like 27-30-33 depending on how you will connect the LL2733 400mH or 100mH.
Using a choke as the first element will slow down the charging of the caps at switching on. BUT big transformers depending on if they are EI, R core or toriodal style will make your light dim for a while when switching on. Big toriods 500VA or more are famous for this. Both my R cores 500VA and 650 VA no problem at all.
In the Technical files of Lundahl you can see that if you use choke input the transformer can deliver more current than used with a cap input. The output current from rectifier with choke input is 95% of the transformer specs with capacitor input it will be 63%. One more good reason to give choke input a try!
Greetings, Eduard
Pano, for what I know there are transients on bass too.
I've read years ago part of that french documentation, I'll read it all. Btw - thanks Sergiu for the link to that Index.
Using coils is a big step forward to a great PS, despite the relatively big cost involved.
A soft start module is cheap and usefull when trafo > 5-600VA. It also permit a low charging of capacitors.
Of course, a high quality relay is mandatory 🙂
I was wrong yesterday recommending lower L with very low Re - it's good for an AB class amp.
In this case of an A class, it is not so important to keep Re very low as long as the consumption is constant, so the supply voltage will not variate.
I've read years ago part of that french documentation, I'll read it all. Btw - thanks Sergiu for the link to that Index.
Using coils is a big step forward to a great PS, despite the relatively big cost involved.
A soft start module is cheap and usefull when trafo > 5-600VA. It also permit a low charging of capacitors.
Of course, a high quality relay is mandatory 🙂
I was wrong yesterday recommending lower L with very low Re - it's good for an AB class amp.
In this case of an A class, it is not so important to keep Re very low as long as the consumption is constant, so the supply voltage will not variate.
Last edited:
A few days ago a good question was asked : can it be used one double coil on a single core (bifilar winding) ? Well, I think it does, and it's even [much] better. Condition to connect the correct beginning and end of each winding.
A new problem occurs then, positioning the coil vs trafo, the coil is sensitive to any fields.
A new problem occurs then, positioning the coil vs trafo, the coil is sensitive to any fields.
Last edited:
Hello,
When doing this the current in both coils should be identical. Probably this will be very hard to be possible.
When using R core power transformer and Lundahl chokes mutual influence is very very low. R core radiates very low and Lundahl does the same. I did do some test with an anode choke from Lundahl and a dnm meter. Anode choke ( 60000mH ) will pick up much more than a 400mH choke. Did run more than 2A through the R core.
Greetings, Eduard
Why ? The current is quasi identical on both + and - rails.
I know that R-Core are very low radiating and LL chokes well shilded, I was referring to chokes and trafos, generally speaking. Not everybody afford RCore and Lundahl.
As I see the problem, trafo is emmiting and the coil is receiving, there is practically no emission from the coil if they are connected 2 windings on a single core.
Mutual influence must be very very low because humm is extremely easy to hear and usually these amps are used on speakers with high (or very high) SPL.
Last edited:
@shambala & Pano:
Just saw this and may add:
... to my knowledge an ideal square wave (front & backside of a "transient") with infinitely low rise time is composed of all frequencies existing from at least close to DC to infinity. A bass transient would have a much lower rise time and thus would have a much lower span of inherent frequencies and typically those dominant would be lower frequencies.
Hope not this is too much out of context ;-)
Cheers,
Jesper
Just saw this and may add:
... to my knowledge an ideal square wave (front & backside of a "transient") with infinitely low rise time is composed of all frequencies existing from at least close to DC to infinity. A bass transient would have a much lower rise time and thus would have a much lower span of inherent frequencies and typically those dominant would be lower frequencies.
Hope not this is too much out of context ;-)
Cheers,
Jesper