That was why I said the imput wave looks triangular. Often this is dismissed as a ghost in the machine. Richard Hay of Nytech advised Julian that great sounding amplifiers have similar current and voltage waveforms.
Member
Joined 2009
Paid Member
The amended FFT is attached.
to my inexperienced eyes, that's a nasty looking amount of 5th harmonic (I suspect the stuff out past 1MHz won't be reliable since the simulation ignores layout and device parasitics).
That looks dreadful. About 20 to 30dB worse than I would expect. Hope you don't get that. The rest is about right.
I have some notes on nap250. 2nd -72dB 3rd -75dB 4th -95dB 5th -87dB. It shows fifth is about 13dB higher than ideal whilst not a disaster. As I have said it's not the ski slope curve they might have been looking for. I speculate a single input transistor version should have that.
to my inexperienced eyes, that's a nasty looking amount of 5th harmonic (I suspect the stuff out past 1MHz won't be reliable since the simulation ignores layout and device parasitics).
Naim recommend the use of their brand of speaker cable - these would be part of the load. I have no information on the properties.
Member
Joined 2009
Paid Member
I have some notes on nap250. 2nd -72dB 3rd -75dB 4th -95dB 5th -87dB. It shows fifth is about 13dB higher than ideal whilst not a disaster. As I have said it's not the ski slope curve they might have been looking for. I speculate a single input transistor version should have that.
look at post 106 for my FFT results from my Spice file.
In the first days RS 50 strand 2.5 mm cable was used twisted 2.5 times per metre. The NACA4 degraded fast due to plasticiser. NACA5 was claimed not to have the plasticiser, feels like polypropylene to me. It's 3/4 inch spacing like international 4mm socket spacing.
All flexible PVCs are plasticised but the older products used a form of the oily substance (di-allyl phthalate) that literally sweated out of the PVC over time/temperature, leaving it somewhat brittle and left a sticky residue wherever the stuff was stored for long periods. Early PVC rainwear and safety clothing were infamous for this, back in the day. Monster Cable had a bad reputation for corroding the copper flex too. As a result, most PVC produced now incorporates a modified form of the plasticiser that is more stable.
Naim's explanation for the change of cables, via their forum, covers for the fact that their usage wouldn't support an economical production run when the main user stopped buying it. NACA4 was an existing product when Naim began supplying it long ago but it was obviously nothing special or unique other than being heavy duty flex. It has reasonably close conductor spacing which can be acceptably matched with similar capacitance and inductance/metre in cheap cables too. NACA5 is actually sheathed in HDPE which has a higher temp. rating than PVC, as required for modern building cables. It's certainly stiffer as a result. Surprising to me, it's not new either. NACA5 was introduced in 1989 according to Martin Colloms: http://www.tomtomaudio.co.uk/reviews/NACA5 speaker cable.pdf
Whatever the case, it's nothing like NACA4 that was claimed to be best at least for early NAP models and all the huff and puff about matching inductance per metre and total capacitance must be nonsense if we accept that the same 3.5m length of either cable will work equally well in a system.
Naim's explanation for the change of cables, via their forum, covers for the fact that their usage wouldn't support an economical production run when the main user stopped buying it. NACA4 was an existing product when Naim began supplying it long ago but it was obviously nothing special or unique other than being heavy duty flex. It has reasonably close conductor spacing which can be acceptably matched with similar capacitance and inductance/metre in cheap cables too. NACA5 is actually sheathed in HDPE which has a higher temp. rating than PVC, as required for modern building cables. It's certainly stiffer as a result. Surprising to me, it's not new either. NACA5 was introduced in 1989 according to Martin Colloms: http://www.tomtomaudio.co.uk/reviews/NACA5 speaker cable.pdf
Whatever the case, it's nothing like NACA4 that was claimed to be best at least for early NAP models and all the huff and puff about matching inductance per metre and total capacitance must be nonsense if we accept that the same 3.5m length of either cable will work equally well in a system.
I tried several cables with my clone and two or three original naim and I got the best results with untwisted cables, with either large strands or Litz 30 micron wire with many strands per conductor (360 and 500) is a cable that a lost friend made with ptfe sheath.
The worst result was with multi-stranded braided cable, it gave a flat sound and with artifacts in the treble.
The worst result was with multi-stranded braided cable, it gave a flat sound and with artifacts in the treble.
Just for the record, The plasticiser I referred to in #1168 was actually Dioctyl phthalate, not that it will mean much here.
To be frank NACA 5 wasn't great. I like DNM cable. No need to spend silly money on it. Mr Westwood told me the 2.5 turns per metre was exactly the Tannoy spec back to their origins. He was my boss when 17 11/12 years old.
Member
Joined 2009
Paid Member
DNM cable is not just cable - at least I remember reading that Dennis puts a kind of zobel on the ends of his cables to terminate them at RF. He has some interesting approaches to all of his products.
Very genuine guy Denis. Notice prices much like audiophile companies now. Suspect they make a Crimson derived amplifier. On the whole Crimson was my favourite amp of the period.
Well, I hope eventually you too find your own ‘TGM8’, that special amp which your ears tells you should not be improved.
At last i have simulated the TGM8. I can see why you think it is special. Unlike the P3A, the TGM8 is indeed a proper amp (sorry, Rod). It is the input topology that makes it possible (actually LTP input is possible too but is difficult and not as good as singleton input; hence the bad reputation of LTP for moderate bandwidth amplifiers). The price to pay is the higher THD (simulated 0.01% at 20V output) and of course low PSRR. I'm curious to find out if the extra distortion worth the extra 'correctness', will build one in a few days.
Why is C16 (100pF) in TGM8?
Oscillation. It can be reduced but TGM8 is already good with 47pf Miller cap.
Member
Joined 2009
Paid Member
And this is where we have to be careful with the Naim circuit because the Miller cap is only one part of the story - the VAS device has high Cob. So now we have a total dominant pole capacitance that varies with the voltages on the VAS device. Given that lower Cob devices were available to JV he had to have made it a conscious decision to use what he did.
If you mean the Ferranti made devices I doubt he gave it much thought. He had a real panic when it was possible they were bankrupt. The big deal was usually much better current, gain and dissipation than any other make. The original designer most likely would have set the parameters. Mostly the Japanese designed using low cob devices. 2SB716 and 2SD756 come to mind. One simply uses a larger vas cap. A high grade capacitor being better than the internal capacitance. COG NPO types usually.
- Status
- Not open for further replies.
- Home
- Amplifiers
- Solid State
- TGM10 - based on NAIM by Julian Vereker