Hi Erin,
I wonder your procedure to modify the XPA-2 would be easy enough for a average handy guy like me.
I am in Sydney and wonder if you could help by giving instructions to modify the xpa-2. I have not ordered one yet, but as my previous post, I am tossing between xpa-1 and xpa-2 to run my Tanoy dc10t speakers.
If I could perform the same modifications as you did than the xpa-2 could be a good option and good for the pocket as well.
I wonder your procedure to modify the XPA-2 would be easy enough for a average handy guy like me.
I am in Sydney and wonder if you could help by giving instructions to modify the xpa-2. I have not ordered one yet, but as my previous post, I am tossing between xpa-1 and xpa-2 to run my Tanoy dc10t speakers.
If I could perform the same modifications as you did than the xpa-2 could be a good option and good for the pocket as well.
I'm about to install some new signal wire in it so I will take a photo for you either today or tomorrow.
Good timing then to ask, thanks! Are you going to replace the cables that go to speaker output terminals (and maybe something else?) and what king of cable do you plan to use?
XPA-1 vs. XPA-2
Legis, I put both photos together on one post... hope you don't mind me using your photo.
As can be seen the XPA-1 has nice thick tracks and nice capacitors.
XPA-2 has thinner tracks and not very nice capacitors.
In practice though, the XPA-2 sounds very nice.
But not as nice as the XPA-1. Yes, I have heard both.
Legis, I put both photos together on one post... hope you don't mind me using your photo.
As can be seen the XPA-1 has nice thick tracks and nice capacitors.
XPA-2 has thinner tracks and not very nice capacitors.
In practice though, the XPA-2 sounds very nice.
But not as nice as the XPA-1. Yes, I have heard both.
Attachments
Last edited:
Thanks erin for the photos! I always thought that XPA-1 is very close to XPA-2 (mainly difference being that XPA-1 is bridged) but it definitely seems that these amps don't haver similar amp modules. Very good and neat job on both board.
By the way, I'm going to change the XPA-1s' emitter resistors to non-inductive Mills, and change the value from 0,5ohms (like XPA-2 also has) to 0,22ohms (and up the wattage rating to 12W) so I get over double A-class region with the same bias voltage that is used originally (approx. 10mV measured over one bias resistor, changing little to resistor to resistor). Also the output coil's damping resistor will be changed to Mills, as it's also directly on the signal path like emitter resistors. The wires that go to speaker terminal will be changed to long grain solid core copper and for the two big 4,7µF DC removing caps at the preamp stage near the inputs, I thought that 250V ESA-series Clarity Caps would be a good choice, mainly because of their size (they are also reasonably prized).
I could also change the two caps on the amp modules (I think they are bypass caps between voltage rails and ground), and one "HF shorting" cap between the voltage rails, but ESA-series does not have right value caps so maybe I just leave them be. Maybe some ultra low loss factored polyprop caps would give better HF filtering though but the change can be very minor. There are also other caps (Wimas) but I have no idea what cap is on the signal path.
By the way, I'm going to change the XPA-1s' emitter resistors to non-inductive Mills, and change the value from 0,5ohms (like XPA-2 also has) to 0,22ohms (and up the wattage rating to 12W) so I get over double A-class region with the same bias voltage that is used originally (approx. 10mV measured over one bias resistor, changing little to resistor to resistor). Also the output coil's damping resistor will be changed to Mills, as it's also directly on the signal path like emitter resistors. The wires that go to speaker terminal will be changed to long grain solid core copper and for the two big 4,7µF DC removing caps at the preamp stage near the inputs, I thought that 250V ESA-series Clarity Caps would be a good choice, mainly because of their size (they are also reasonably prized).
I could also change the two caps on the amp modules (I think they are bypass caps between voltage rails and ground), and one "HF shorting" cap between the voltage rails, but ESA-series does not have right value caps so maybe I just leave them be. Maybe some ultra low loss factored polyprop caps would give better HF filtering though but the change can be very minor. There are also other caps (Wimas) but I have no idea what cap is on the signal path.
Last edited:
Hi Legis,
I have been interested in the sound of emmiter resistors. Have you changed to non inductive types before? Does it make a big difference to the sound?
Will you be using 1mm diameter solid core copper, and paralleling wires, or something thicker?
I'm sure you will notice any upgrades to your XPA-1, but I presume you like the sound of them stock?
I really wish I could get the schematic to my amp!!
I have been interested in the sound of emmiter resistors. Have you changed to non inductive types before? Does it make a big difference to the sound?
Will you be using 1mm diameter solid core copper, and paralleling wires, or something thicker?
I'm sure you will notice any upgrades to your XPA-1, but I presume you like the sound of them stock?
I really wish I could get the schematic to my amp!!
I notice that both the XPA-1 and XPA-2 have trim pots on the output stage. Do you think this is for DC offset or idle current bias? One channel of my amp idles hotter than the other. I would like to match the bias on both channels. I guess I will have to experiment to find out. The reason I have not done this allready is that the wiring is quite short and it is difficult to get the output stage out, whilst still being connected, and also measure bias, and turn trimmers..... and not short something.....
Hi erin, I have not changed emitter resistors before. I would not expect a huge difference but there are 24pc per channel so all the smallest differences start to accumulate I quess.
I'm going to parallel 19 gauge solid core, maybe 4 wires per terminal to give ~2,6mm2/13ga total thickness. I was thinking Wonder Wire as it's high purity, long grain silver plated copper and has low loss insulation (some kind of polypropylene according to TRT's website).
I have liked stock XPA-1s very much indeed, I have had them for almost a year. They are highly resolving, very musical with expansive and stable soundstage and clear and smooth upper end. Timing and bass control is also nice
. Capability to drive (maybe) every speaker know to man is also a plus.Here is THD plot of stock XPA-1 taken approx. after 2 weeks of ownerhip. 1,3V (sound card's maximum input without voltage divider) into 4,5ohms resistor network dummy load that has 3,6mH of inductance (might be better at HF if the dummy load would be non-inductive):
An externally hosted image should be here but it was not working when we last tested it.
Here's the same configuration (excep higher sample rate) THD plot with emitter resistor's resistance halved by soldering another 0,5ohm resistor accross the original resistor and bias voltage set to original ~10mV:
An externally hosted image should be here but it was not working when we last tested it.
THD has dropped 8dBs at the low freqs and 4dB at HF with the increased bias, while the stock performance being very good also. However I have not had time to listen to the amps yet
(I bet they don't sound worse), they are opened up and waiting for the parts.I notice that both the XPA-1 and XPA-2 have trim pots on the output stage. Do you think this is for DC offset or idle current bias? One channel of my amp idles hotter than the other. I would like to match the bias on both channels. I guess I will have to experiment to find out. The reason I have not done this allready is that the wiring is quite short and it is difficult to get the output stage out, whilst still being connected, and also measure bias, and turn trimmers..... and not short something.....
That's bias voltage pot in XPA-1, I would be inclined to think that it has same fuction also in XPA-2. I have tried to change the bias voltage but found out via measurements that the stock setting of XPA-1 (approx. 10mV voltage drop per emitter resistor) is optimal, regardless the usually recommended 15-25mV. Higher bias rises H3 in A-class region and the amp produces more odd order harmonics also at higher power (50-150W @ 4ohm). For this reason I will change the emitter resistors to smaller ones, to be able to increase the bias current without changing the bias voltage.
If you just want to even the bias balance between the channels there's no harm done for sure. If you have measuring equipment you could try different bias settings, it is not sure if XPA-2 is optimally biased. Have you been able to measure it's power consumption in idle?
Last edited:
Legis, I have tried silver plated copper wire before with DIY interconnect cable. I don't like it. It seems to give an unnatural brightness to the sound. But, perhaps you have had a different experience?
I have not measured its idle consumption across the emitter resistors. I think it may be more tricky with the XPA-2. there is not much room to work with the output stage out.
But since you suggest the trim pot is for bias, I will try to give it a go. As I am sure one channel is lower than the other.
I am also now encouraged to replace the emitter resistors with 0.22R
Are you suggesting I need to increase them from 5 Watt to 10 Watt resistors ?
Thanks
I have not measured its idle consumption across the emitter resistors. I think it may be more tricky with the XPA-2. there is not much room to work with the output stage out.
But since you suggest the trim pot is for bias, I will try to give it a go. As I am sure one channel is lower than the other.
I am also now encouraged to replace the emitter resistors with 0.22R
Are you suggesting I need to increase them from 5 Watt to 10 Watt resistors ?
Thanks
Since the amp power will be the same , 0.22R resistors will see
half of the voltage compared with 0.5R , hence they will dissipate
half of the previous emitter resistors values, so 5W models suits perfectly.
I have had good experinces with silver plated pure copper wires with good insulation, these all variables make a difference. I found that polyester and PCV are not good sounding insulators in cables, just like polyester capacitors are not good either: Measured Differences Between Capacitors for Audio Applications. For example teflon, polyethylene, polypropylene and of course AIR are good insulation materials, air is the best. Every material that has low loss factor and low capacitive effect, is usually good sounding insulator. Also long grain copper (OCC) tends to relax the sound compared to normal copper. I also usually like the sound of solid core over multi-stranded. I have Nordost heimdall speaker cables that I have liked very much, they also have all these variables "correct" to my liking (air/teflon inslulation, solid core, high purity copper). Silver plating usually keeps the conductor unoxidized, as silver does not oxidize (or oxidizes very slow) in room temperature.
Thanks wahab, I though that the dissipation is doubled when the resistance is halved... I though that the voltage would stay the same regardless the resistance, and the maximum voltage every resistor would see, is the amplifer's maximum output voltage divided by emitter resistor count. Halving the emitter resistor's resistance did however did halve the bias voltage accross the resistors if the bias pots were kept at the same position.
I quess I will then take the 5W Mills instead of the 12W.
Last edited:
I ended up removing the Klotz cable that I installed inside the amp and reinstalled the original cable as supplied by Emotiva. The original cable sounds very nice.
I also did one mod. The original audio signal wire runs underneath the power supply PCB through the same hole in the chassis parallel with the speaker wires and power supply wires. I drilled two holes in the top of the side plates so I can run the signal wire on top of the power supply PCB. It means the signal wire is shorter and separated from other wires. I think this sounds better than the original way.
I also did one mod. The original audio signal wire runs underneath the power supply PCB through the same hole in the chassis parallel with the speaker wires and power supply wires. I drilled two holes in the top of the side plates so I can run the signal wire on top of the power supply PCB. It means the signal wire is shorter and separated from other wires. I think this sounds better than the original way.
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
Today I received all the modding incredients (which went through little change):
- 60pcs of Mills wire wound resisitors for audio
- 30ft of UP-OCC hook up wire (14ga and 16ga) from Neotech
- handfull of Clarity caps
I decided to get 0,15R Mills for emitter resistors as the 0,18R were not in stock. I have also read that the smaller the value the better the crossover distortion behavior (beneficial to decrease the value as long as the bias stability is maintained).
By the way, if anyone's interested I measured inductance of various resistors that I had in hand.
- Emotiva's original wire wound emitter resistors 0,5ohm: 0,148µH
- Random wire wound resistor from local electronics shop 0,47ohm: 0,266µH
- Emotiva's axial metal oxide resistors, 10ohm: 2,182µH
- Mills MRA-5 0,15ohm: 0,038µH.
- Mills MRA-5 10ohm: 2,213µH
- Mills MRA-12 1ohm: 0,238µH.
0,15R MRA-5s have very low inductance, one can say it's negligible (at least when there are multiple resistors parallel in the output). The 10 ohm Mills' inductance is almost as low as the metal oxide resistor's. It would be interesting to see how low inductance some audio grade metal film resistors can achiece in higher ohm-values. Wire wound resistors have the lowest noise (they produce only heat related noise), but not the lowest inductance.
Measurements were done by using this meter: Update high precision L/C meter from Ebay
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.