We actually made a sound system for a club some years ago, using 49 pcs 8" (cheap) hifi woofers for bass. Each line of 7 bass'es were connected in parallel to give 1.2 Ohms impedance. Each line was then run on a Millennium amplifier on a +/- 30 V rail.
We used this approach because we want to offer the customer 5 years total (no matter what...) warranty on his new system. In other systems we have used JBL and Beyma 18" woofers in backload horns etc, like a PA engineer would 'usually' do. However we didn't feel this conventional setup was very reliable. We had seen systems fail after a few years of hard beating.
Now the interesting thing about the 49 small (and cheap) speakers, each in a small closed cavity of just 4 liters, is that each unit can take up to 50 W of power safely for years. 49 x 50 is
2450 Watts. Now which set of 18" units can take this kind of beating for extended periods? And further the amplifiers were simpler to build, because only a low rail voltage is required.
Another advantage is that the depth of the entire bass speaker is only 20 cm, making it easy to fit over the dancefloor.
The performance of the system was so great that it was talked about in DJ circles all over the country for years, until now i have not heard any conventional 18" horn that can match the tight sound or even the sound pressure of this special speaker!!! When it's going loose you can actually see your jacket sleeves moving in the sound pressure! We measured 147 dB SPL @ 1M
And reliability ..? Well now 10 years later we have had only one service visit to the system, and that was because a defective ice maker on the same power line had killed a bridge rectifier in one amp. After that 2$ part replacement, it is still playing day after day.
So i think sometimes it is worth thinking 'outside the box'. The 'usual' solution is not necessarily the only good way to go.
We used this approach because we want to offer the customer 5 years total (no matter what...) warranty on his new system. In other systems we have used JBL and Beyma 18" woofers in backload horns etc, like a PA engineer would 'usually' do. However we didn't feel this conventional setup was very reliable. We had seen systems fail after a few years of hard beating.
Now the interesting thing about the 49 small (and cheap) speakers, each in a small closed cavity of just 4 liters, is that each unit can take up to 50 W of power safely for years. 49 x 50 is
2450 Watts. Now which set of 18" units can take this kind of beating for extended periods? And further the amplifiers were simpler to build, because only a low rail voltage is required.
Another advantage is that the depth of the entire bass speaker is only 20 cm, making it easy to fit over the dancefloor.
The performance of the system was so great that it was talked about in DJ circles all over the country for years, until now i have not heard any conventional 18" horn that can match the tight sound or even the sound pressure of this special speaker!!! When it's going loose you can actually see your jacket sleeves moving in the sound pressure! We measured 147 dB SPL @ 1M
And reliability ..? Well now 10 years later we have had only one service visit to the system, and that was because a defective ice maker on the same power line had killed a bridge rectifier in one amp. After that 2$ part replacement, it is still playing day after day.
So i think sometimes it is worth thinking 'outside the box'. The 'usual' solution is not necessarily the only good way to go.
It reminds me of a trick question.
I was demonstrated a single 4" midbass driver. It produces only midbass, no low bass and no trebles.
I was asked what will be the sound like if we put 50 of this? Only midbass or what? The answer is full range. Where does the low bass and trebles coming from? Each driver seems cannot reproduce it, but when they gathered in so many number, it becomes full range.
I was demonstrated a single 4" midbass driver. It produces only midbass, no low bass and no trebles.
I was asked what will be the sound like if we put 50 of this? Only midbass or what? The answer is full range. Where does the low bass and trebles coming from? Each driver seems cannot reproduce it, but when they gathered in so many number, it becomes full range.
I hope we all come back to the ZETA , Nobody answered my last post !!!
All were so busy off topic
Cheers
All were so busy off topic
Cheers
RL: The optimal value depends on the N-channel MOSFET's you use, and how much dissipation you can tolerate in the P channel FET. But i would say for a good compromise use the 47 Ohms suggested. As you can see i used 22 Ohms in the test board, and it works fine too.
rajeev luthra said:
Hi Rajeev,
Kindly give us your Address and phone number so that we can contact you for delivery of PCB .
Email us at amp_man_1@hotmail.com
Regards
Lars,
Are R30-R41 what is being referenced in your post <http://www.diyaudio.com/forums/showthread.php?postid=530205#post530205> that states:
The optimal value depends on the N-channel MOSFET's you use, and how much dissipation you can tolerate in the P channel FET. But i would say for a good compromise use the 47 Ohms suggested. As you can see i used 22 Ohms in the test board, and it works fine too.
John L. Males
Willowdale, Ontario
Canada
10 December 2004 10:51
Are R30-R41 what is being referenced in your post <http://www.diyaudio.com/forums/showthread.php?postid=530205#post530205> that states:
The optimal value depends on the N-channel MOSFET's you use, and how much dissipation you can tolerate in the P channel FET. But i would say for a good compromise use the 47 Ohms suggested. As you can see i used 22 Ohms in the test board, and it works fine too.
John L. Males
Willowdale, Ontario
Canada
10 December 2004 10:51
Keypunch: Of course not! The issue of these source resistors i think has been covered pretty thoroughly previously in this thread.
The resistor referred to is the one going from Drain on the P-channel FET to V- rail.
🙂
The resistor referred to is the one going from Drain on the P-channel FET to V- rail.
🙂
Resistor Going from Drain on the P-channel FET to V- rail
Lars,
With respect to your reply:
<http://www.diyaudio.com/forums/showthread.php?postid=530459#post530459>
Please calm down. I was not asking about the source resistors. My question was not about the source resistors, but to clarify for me (who is not an amp designer) what component was being discussed. I know it has been discussed much in this thread and I know these source resistors have been subject of much discussion in other threads on other mosfet based amplifier designs. That was not the question being asked.
Before posting my question of clarification I had searched high and low for the resistor in question. in fact I thought if might be in reference to the IRFP9240. If the IRFP9240 is the "P-chanel FET" you refer to in your reply "The resistor referred to is the one going from Drain on the P-channel FET to V- rail." please confirm.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 13:47
Lars,
With respect to your reply:
<http://www.diyaudio.com/forums/showthread.php?postid=530459#post530459>
Please calm down. I was not asking about the source resistors. My question was not about the source resistors, but to clarify for me (who is not an amp designer) what component was being discussed. I know it has been discussed much in this thread and I know these source resistors have been subject of much discussion in other threads on other mosfet based amplifier designs. That was not the question being asked.
Before posting my question of clarification I had searched high and low for the resistor in question. in fact I thought if might be in reference to the IRFP9240. If the IRFP9240 is the "P-chanel FET" you refer to in your reply "The resistor referred to is the one going from Drain on the P-channel FET to V- rail." please confirm.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 13:47
Schematic Correction for Resistor Going from Drain on the P-channel FET to V- rail
Lars,
Thanks for your reply. Based on your reply:
<http://www.diyaudio.com/forums/showthread.php?postid=530516#post530516>
it would appear to me the currently posted version (22 Sept 2004) of the Zeta schematic that the resistor in question if from drain of T14 to the -supply rail line. Currently the resistor in question os labled "10", implying it is 10 ohms and has no R part number assigned to it in the schematic. Would be correct to deduce from your reply above? If so, can you see why someone who is trying to follow what is being discussed may have asked the question to clarify and in terms of the schematic part number references?
I understand you are an engineer and relate to the purpose, stage, etc of a design and not the part number on the schematic. Please understand some of us are not so knowledgeable of the finer points of design, circuit functions or ability to read minds that we will ask such questions so we do not fail or maybe blow something up. This might lead to possible injury or cost of parts.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 16:01
Lars,
Thanks for your reply. Based on your reply:
<http://www.diyaudio.com/forums/showthread.php?postid=530516#post530516>
it would appear to me the currently posted version (22 Sept 2004) of the Zeta schematic that the resistor in question if from drain of T14 to the -supply rail line. Currently the resistor in question os labled "10", implying it is 10 ohms and has no R part number assigned to it in the schematic. Would be correct to deduce from your reply above? If so, can you see why someone who is trying to follow what is being discussed may have asked the question to clarify and in terms of the schematic part number references?
I understand you are an engineer and relate to the purpose, stage, etc of a design and not the part number on the schematic. Please understand some of us are not so knowledgeable of the finer points of design, circuit functions or ability to read minds that we will ask such questions so we do not fail or maybe blow something up. This might lead to possible injury or cost of parts.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 16:01
keypunch: I understand your point. However as you might have noticed the amplifier is already being built in several different versions. I was discussing with Rajeev Luthra about his version with 2 pairs of IRFP250 (not 5 - 10 pairs as i have proposed in the published schematics). Also Rajeev is using another MOSFET than proposed. For his purpose (since he aksed me the question) i have given him the best answer.
So don't let yourself be confused with the change of values from the schematic. If you build it exactly after the schematic you should still just use the values and parts mentioned on the schematic. Just Rajeev (who asked me the question you are referring to) built another version than printed on the schematic.
Hope this clairifies all. 🙂
BTW the schematic has been downloaded almost 4000 times by now, i hope more people will soon write about their experiences with the Zeta project. Thank You 🙂
So don't let yourself be confused with the change of values from the schematic. If you build it exactly after the schematic you should still just use the values and parts mentioned on the schematic. Just Rajeev (who asked me the question you are referring to) built another version than printed on the schematic.
Hope this clairifies all. 🙂
BTW the schematic has been downloaded almost 4000 times by now, i hope more people will soon write about their experiences with the Zeta project. Thank You 🙂
Hi Lars,
Thanks for your reply. I have been following the thread from the start. I guess in all the does this varient or change work or ok, I did not catch onto Rajeev my question being related to a variation on the theme.
I have an interest in this amp for a couple different reasons. The pros being core transistors I can get and seems design is flexible enough to accomodate alternate IRF output devices to suit my application requirements as Rajeev has done, and seems along lines of the varients I have been thinking and previous inquired about.
The cons are some mention of the design being more of PA quality. I do not need super high fi, but I have a desire to use the same amp and output devices if I decide to bi or tri amp later. For low frequencies I am willing to have a couple paralleled output devices. My reading about this fourm and other sources suggests for mid and high range frequencies parallel output devices tend to seriously affect such signals which is of no concern for low fewquencies. Although I am not a engineer, I also have concern T14, the IRFP9240 is a non-mirror effect on the signal. You do not need to give me thory or otherwise on this T14 point. This is just a plain old non-engineer thinking person opinion.
Opinion aside, I still follow the design and thread to enable me to evaluate my opinion on ongoing basis.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 19:26
Thanks for your reply. I have been following the thread from the start. I guess in all the does this varient or change work or ok, I did not catch onto Rajeev my question being related to a variation on the theme.
I have an interest in this amp for a couple different reasons. The pros being core transistors I can get and seems design is flexible enough to accomodate alternate IRF output devices to suit my application requirements as Rajeev has done, and seems along lines of the varients I have been thinking and previous inquired about.
The cons are some mention of the design being more of PA quality. I do not need super high fi, but I have a desire to use the same amp and output devices if I decide to bi or tri amp later. For low frequencies I am willing to have a couple paralleled output devices. My reading about this fourm and other sources suggests for mid and high range frequencies parallel output devices tend to seriously affect such signals which is of no concern for low fewquencies. Although I am not a engineer, I also have concern T14, the IRFP9240 is a non-mirror effect on the signal. You do not need to give me thory or otherwise on this T14 point. This is just a plain old non-engineer thinking person opinion.
Opinion aside, I still follow the design and thread to enable me to evaluate my opinion on ongoing basis.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 December 2004 19:26
Hi Lars,
I have increased the value of R12 to 3k3 in my Zeta amp and now its
working ok , I think this was the problem .
Yesterday on turn ON two Irfp250 one from each rail failed ie ( the drain
source are showing short ) , I disconnected these and the amp is
working fine , I do not understand WHY two mosfets failed , I had used
closely matched mosfets two in each rail out of which one from each rail
failed .
I would like to know if it is Ok to use only two mosfets in the output , with
a 8 ohm load and a 40v-0-40v transformer .
Thanks
Rajeev
I have increased the value of R12 to 3k3 in my Zeta amp and now its
working ok , I think this was the problem .
Yesterday on turn ON two Irfp250 one from each rail failed ie ( the drain
source are showing short ) , I disconnected these and the amp is
working fine , I do not understand WHY two mosfets failed , I had used
closely matched mosfets two in each rail out of which one from each rail
failed .
I would like to know if it is Ok to use only two mosfets in the output , with
a 8 ohm load and a 40v-0-40v transformer .
Thanks
Rajeev
If you are using genuine IR parts IRFP250 then a single pair of these would be OK for 8 ohm load up to the 56V rails you would have with 40-0-40 transformer.
Note that IRFP250N is not as robust as IRFP250 as the die is smaller. I don't know how well the N version would cope in this situation.
Note that IRFP250N is not as robust as IRFP250 as the die is smaller. I don't know how well the N version would cope in this situation.
Just checked IRFP250N datasheet and it seems about the same SOA as IRFP240, so should be OK in this case. You would be better off with the IRFP240 though as the gate capacitance and charge are much lower.
IRFP250 is I believe basically a pair of IRFP240 dies in one case and would be fine for 4 ohm load at 56V rails.
IRFP250 is I believe basically a pair of IRFP240 dies in one case and would be fine for 4 ohm load at 56V rails.
Keypunch: In case you want to optimize sound quality at cost of reducing power output (from 900 W to 200 W) you should make a single MOSFET setup with the Zeta circuit. Then you are away from 'PA sound' and on the way to high performance Audio sound.
A well known danish company make excellent high-end amplifiers (very expensive too) with Single pair n-Channel MOSFET topology, just like the Zeta would then be. Try the search phrase:
"Single Mosfet Output" Audio
on Google, then you should see more.
A well known danish company make excellent high-end amplifiers (very expensive too) with Single pair n-Channel MOSFET topology, just like the Zeta would then be. Try the search phrase:
"Single Mosfet Output" Audio
on Google, then you should see more.
Lars Clausen said:
But is that REALLY the case? Are those MOSFETs not actually the standard type for that kind i.e. a number of separate power chips in one package and thus really more than one MOSFET and the statement about using one single MOSFET only marketing? I thought they were using these:
http://www.magnatec-uk.com/pdf/magnatec/BUZ900X4.pdf
Hi Lars,
I have no need for power beyond 150W/8 ohms as far as I can determine from my own listening needs. So for sure if I can obtain 200W/400W/800W in 8/4/2 ohms I am more than fine when using speakers with passive crossovers. Please note the 2 ohm rating is for safety margin. As I see it, as minimum I need 100W/200W/400W when in active crossover to woofer, 60W/120W/240W for midrange, and 20W/40W/80W for tweeter in same active based system using 24DB L-R based design.
With respect to:
A well known danish company make excellent high-end amplifiers (very expensive too) with Single pair n-Channel MOSFET topology, just like the Zeta would then be. Try the search phrase:
"Single Mosfet Output" Audio
I tried the search even though I was certain what the intended result was to be. I did not obtain the intended result 🙁 Assuming I am correct, it would be far beyond my budget for single stereo unit, let alone 4 for passive based system, and the 20 I need for active based 3 way, or 14 for active based 2 way. Assuming the same amp thinking of, too many only available SMD which seems to too complicated to solder by hand, let alone PCB I need to rework should I choose to reduce the number of SMD to increase success. For sure the design and circuit interest me, but far from my budget abilities.
UrSv,
Hi, I think we had a few offline eMails few days ago.
With respect to if the single mosfet vs paralleled mosfets, I have in fact found a very large pool of thinking and reasoning why parallel mosfet output driver stages are inferior.
I fully realize there are many well regarded designs that in fact use paralleled output devices. Many of the Pass amps in fact do use parallel output devices. So like all things in Audio, execution as well as theory decide final outcome.
Does that mean some comapnies will try to peddle the single mosfet as better when their designs are inferior to parallel based designs? Not likely. That being case, it is up to the consumer to be informed to make the decision on fact not hype. We all know why marketing exists 😉
For me the single pair output device is of interest based mostly the theory and practical issues that surrround paralleled output mosfet devices. I am willing to consider 2 or 3 pairs of paralleled mosfet output devices for the woofer side of active system as it seems generally accepted the issues with paralleled output drivers impacts by and lagre the midrange and highs, not the low frequency range as I understand from my research reading.
Regards,
John L. Males
Willowdale, Ontario
Canada
12 December 2004 14:45
I have no need for power beyond 150W/8 ohms as far as I can determine from my own listening needs. So for sure if I can obtain 200W/400W/800W in 8/4/2 ohms I am more than fine when using speakers with passive crossovers. Please note the 2 ohm rating is for safety margin. As I see it, as minimum I need 100W/200W/400W when in active crossover to woofer, 60W/120W/240W for midrange, and 20W/40W/80W for tweeter in same active based system using 24DB L-R based design.
With respect to:
A well known danish company make excellent high-end amplifiers (very expensive too) with Single pair n-Channel MOSFET topology, just like the Zeta would then be. Try the search phrase:
"Single Mosfet Output" Audio
I tried the search even though I was certain what the intended result was to be. I did not obtain the intended result 🙁 Assuming I am correct, it would be far beyond my budget for single stereo unit, let alone 4 for passive based system, and the 20 I need for active based 3 way, or 14 for active based 2 way. Assuming the same amp thinking of, too many only available SMD which seems to too complicated to solder by hand, let alone PCB I need to rework should I choose to reduce the number of SMD to increase success. For sure the design and circuit interest me, but far from my budget abilities.
UrSv,
Hi, I think we had a few offline eMails few days ago.
With respect to if the single mosfet vs paralleled mosfets, I have in fact found a very large pool of thinking and reasoning why parallel mosfet output driver stages are inferior.
I fully realize there are many well regarded designs that in fact use paralleled output devices. Many of the Pass amps in fact do use parallel output devices. So like all things in Audio, execution as well as theory decide final outcome.
Does that mean some comapnies will try to peddle the single mosfet as better when their designs are inferior to parallel based designs? Not likely. That being case, it is up to the consumer to be informed to make the decision on fact not hype. We all know why marketing exists 😉
For me the single pair output device is of interest based mostly the theory and practical issues that surrround paralleled output mosfet devices. I am willing to consider 2 or 3 pairs of paralleled mosfet output devices for the woofer side of active system as it seems generally accepted the issues with paralleled output drivers impacts by and lagre the midrange and highs, not the low frequency range as I understand from my research reading.
Regards,
John L. Males
Willowdale, Ontario
Canada
12 December 2004 14:45
John,
We did exchange some e-mails.
I hear you and my post was just to point out that the "single" MOSFET perhaps was more marketing than actual technical superiority by using ONE MOSFET. I fully support the view that one would be preferable and by choosing one big real single MOSFET which can handle the powers you need would perhaps be a good thing rather than parallelling many smaller.
We did exchange some e-mails.
I hear you and my post was just to point out that the "single" MOSFET perhaps was more marketing than actual technical superiority by using ONE MOSFET. I fully support the view that one would be preferable and by choosing one big real single MOSFET which can handle the powers you need would perhaps be a good thing rather than parallelling many smaller.
I can't say about them.
However if you build the Zeta with a single pair of IRFP250N i can GUARANTEE you that you will get TRUE single MOSFET performance 😉
However if you build the Zeta with a single pair of IRFP250N i can GUARANTEE you that you will get TRUE single MOSFET performance 😉