Member
Joined 2009
Paid Member
No, my employer got out of the speaker business before I joined them. And AR never made vacuum tubes. Try another guess 🙂
It seems that most of the AR speakers are sealed cabinets. I have the 66BX three way's . Don't see a lot of them. Why do people prefer ported to sealed speakers ??
Just a thought!🙂
Just a thought!🙂
Very good question. Since a while and also due to the resurrection of the ARC loudspeakers (in which I am involved on consultant base) I completely stopped to calculate or make bassreflex speakers any more. I have listened to this specific bass quality long enough.
Naim loudspeakers my have been the most discussed items they have ever produced, loved by some and hated by others. But I do not remember any bassreflex speakers, they were either closed box (IBL, NBL, DBL, Ovator) or controlled loss (SBL, SL2). The smaller ones may have been different.
It must have been on purpose, since almost all the other manufacuturers made bassreflex loudspeakers, producing that sound almost everybody seems to accept as "good bass sound" (if properly done). IMHO this is simply wrong. Many chassis manufacturers even stopped to make chassis with the right specs for resonable sized closed box designs.
Sorry to be off topic, but it may be of some significance in this thread.
Getting back on topic and talking of Naim amplifier clones if you add cascodes etc it will not to put it simply be a Naim clone!
I wonder what a clone is?
If you walked into the Naim factory collected all the parts and assembled it at home would it be a clone ???
Is a 1985 unit the same as a 1995 unit I wonder
Trev
I wonder what a clone is?
If you walked into the Naim factory collected all the parts and assembled it at home would it be a clone ???
Is a 1985 unit the same as a 1995 unit I wonder
Trev
It would be OK and is almost the Nait 1. However if BC560C+BF721 and BF720 CCS we could look to at least double the voltage of the now amplifier. Thus to go to 45 V rail would be no problem. The input pair should be 65V types if so.
If I was doing a production run of PCB the Cascode is future proof and almost certainly better. To have a sweeter and more open Naim type idea is not too bad. It's tonal colour would be much the same.
BC560C allows a super high gain and super fast device drive the BF721 in common base. This will get every last drop of speed out of the circuit. The input Z of the BC560C might be 1K which is more ideal ( ZXT753 circa 400R ). That means more current to drive Cdom. The resistor used in the example was 50R ( 51R a standard value, even 10R should work, 33R is about right ). If 10 mA that is 0.5 V loss, Often 5 V total loss is required. The resistor is inserted to avoid Early effect, If you take the internal resistance to be 2R5 of the BF721 the 51R swamps it. If not the transistor is trying to opperate just after conduction starts ( 0.55V typical ). The distortion it makes is second harmonic. One has to be sure you want to get rid of that. The standard Naim VAS will cause plenty of second harmonic. If you look at Self's example it is very easy to do. I used a LED to bias mine. That was a Self type without " Magic " resistor.
Naim fought at the European court to not have to follow directives on rectifier RFi. Often a 10 nF to each diode is the crude type. Naim said fitting them ruined the sound. One can fit soft recovery diodes to make a good compromise. Other make equipment often didn't like Naim amps for this reason. One 10 nF to the AC was often the compromise of other companies.
I think with enough understanding it should be a good move. I was very surprised how well behaved Cascodes are ( in valves also ).
No 688 seems to go too far and less ideal than the simple clone.
I was to meet Ed Vilchur the JV of AR. Alas my friend who was to take me died. Ed and myself both lived in a town of the same name. I was told " No BS Nigel , he won't take it ". I don't think I would have been tempted. When Ed was asked why his turntabels " image so well " he got very upset. I wish I had been there. It is because the movement of the suspension is vertical which will enhance stereo. No magic, just magic at the price. The AR speakers of the pre 1967 era were very different and mostly still are. IB is the poormans version.
http://www.diyaudio.com/forums/solid-state/256552-improvement-vas.html#post3933195
The more I think about it the second example shown here is ideal. On reflection I want to keep it simple so as to keep the tonal colour. If you see we have made a new "super" transistor for want of one resistor and one zener diode. Cdom is as before except it should be easier to drive. Although a unlikely problem stability needs to be checked. There is nothing to stop us using BC560C and MPSA92. 10mA at 45V is just about OK. Make a small TO92 heat sink perhaps.
My brother fixed up an old Goodmans 80 receiver. The VAS emitter was raised up on a diode. This gave 1.2 V at the VAS base and was fed from 2K4. This set the current at 0.5 mA. That all looks very nice until one calculates Zin 250 ohms. The understanding as now is it is current drive so of no importance. If a very low resistance in inserted in the base the real distortion can be read. It could be the track resistance which would be ideal. If one measures the voltage waveform it is like a highly distorted triangle wave. We are told to ignor this.
The Goodmans for once actually had IM distortion and sounded coarse so wasn't in the rhelm of imagination. First attempt was a higher gain transistor. Resulted in oscillation. The standard transistor ( gain 100 ) was put back in. This time the diode was replaced by 56R which gave the same base voltage. Now Z in was 5K6. That's very nice. And guess what, it sounded nice and IM distortion was gone. Loop gain was reduced no doubt. If we wanted to ride both horses at once we might have been able to shunt the 56R with 330nF. If stability was OK that would increse loop gain at the brake frequency which is often 10kHz typical. If stable that would be a by ears test. When completed the Goodmans was a really nice amp that in some ways was better than the Naims of the day. The other updates were capacitors which were best of type from RS ( not audiofool types ). None of you here would have liked the standard version. I suspect all would have like the better VAS version.
Adding 10R to the Naim VAS emitter might help and go to 33K-1K if so. Not sure I would add a shunt. The base would shift 0.1V. 10R in some ways would exceed the Cascode if the simple type, even 33R if wanting a Goodmans result. My big doubt when ZTX753 is the internal capacitance. I think memory distortion is real and this might help.
The more I think about it the second example shown here is ideal. On reflection I want to keep it simple so as to keep the tonal colour. If you see we have made a new "super" transistor for want of one resistor and one zener diode. Cdom is as before except it should be easier to drive. Although a unlikely problem stability needs to be checked. There is nothing to stop us using BC560C and MPSA92. 10mA at 45V is just about OK. Make a small TO92 heat sink perhaps.
My brother fixed up an old Goodmans 80 receiver. The VAS emitter was raised up on a diode. This gave 1.2 V at the VAS base and was fed from 2K4. This set the current at 0.5 mA. That all looks very nice until one calculates Zin 250 ohms. The understanding as now is it is current drive so of no importance. If a very low resistance in inserted in the base the real distortion can be read. It could be the track resistance which would be ideal. If one measures the voltage waveform it is like a highly distorted triangle wave. We are told to ignor this.
The Goodmans for once actually had IM distortion and sounded coarse so wasn't in the rhelm of imagination. First attempt was a higher gain transistor. Resulted in oscillation. The standard transistor ( gain 100 ) was put back in. This time the diode was replaced by 56R which gave the same base voltage. Now Z in was 5K6. That's very nice. And guess what, it sounded nice and IM distortion was gone. Loop gain was reduced no doubt. If we wanted to ride both horses at once we might have been able to shunt the 56R with 330nF. If stability was OK that would increse loop gain at the brake frequency which is often 10kHz typical. If stable that would be a by ears test. When completed the Goodmans was a really nice amp that in some ways was better than the Naims of the day. The other updates were capacitors which were best of type from RS ( not audiofool types ). None of you here would have liked the standard version. I suspect all would have like the better VAS version.
Adding 10R to the Naim VAS emitter might help and go to 33K-1K if so. Not sure I would add a shunt. The base would shift 0.1V. 10R in some ways would exceed the Cascode if the simple type, even 33R if wanting a Goodmans result. My big doubt when ZTX753 is the internal capacitance. I think memory distortion is real and this might help.
Last edited:
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
Later versions of the Naim circuit had a 5R6 emitter resistor with the ZTX. The CCS current set resistor was also increased from 62R to 68R which would have some slight impacts on gain and internal capacitance.
I had hoped so too, until I realised that probably all those similar Amega microprocessor based testers give coarse incremental readings. That rather ruins their usefulness for matching. 'Still fine for ball-park figures and general testing though
Member
Joined 2009
Paid Member
Ah, now I get it. I use that sort of rationale when I've already paid the money too 😉 Next time you have a few quid and in the UK though, a DC75 will really please and the graphics software will entertain for many hours. Peak Electronic Design Limited - Atlas DCA Pro - Advanced Semiconductor Analyser - Model DCA75
Anyway, how is the build progressing? 🙂
Member
Joined 2009
Paid Member
I only spent 10 bucks, that isnt bad if just for Curiosity to see the thing work - i like how it finds the pin outs of various devices. But it doesnt do the nice curve tracing of your Atlas gadget.
Build isnt progressing, the day job and kids and preparing year end taxes is sucking the life out of my hobby. I will make a concerted effort to order the parts this weekend, theyll arrive quickly
Build isnt progressing, the day job and kids and preparing year end taxes is sucking the life out of my hobby. I will make a concerted effort to order the parts this weekend, theyll arrive quickly
Last edited:
There are many "clones" of this topology yet they don't have similar "tonal colour", do they? The tonal colour comes from several things beside the topology itself. And I haven't seen a clone that is a correct (no wonder I have never liked what I heard) interpretation of Naim amp (I only studied NAP250 but I believe all of JV amps have the same thing).
If changing one transistor will change the Naim sound, changing topology (adding a cascode) will make a new different amplifier! It doesn't mean that it cannot be made to have similar tonal colour, but if the design objective is not clear, the tonal colour objective will not be achieved.
As I see it, it is not how big or small the Cdom. They are only part of the big picture. In the Naim amp, is the ZTX plus 39pF is an issue for the LTP? My prediction is not because we have seen worse situation in other amps.
From my short study of this topology to create another clone, I can use MPSA42/92 (even 2N5401), but on simulation I prefer Cdom=47pF, slightly higher than what JV used with the ZTX.
Your new VAS idea, it is probably not hard to drive, but theoretically it has less capability to drive. Yes we have a driver here, but is it sufficient? I think that is the question to find out why such topology usually doesn't sound good.
I found out that this Naim topology will benefit from lower negative rail than the positive rail. I plan to check my transformer first (I think I have several transformers with secondaries of different windings) then I will design my first clone based on this transformer. That could be a genuine improvement.
- Status
- Not open for further replies.