Sansui 800 Resto - she played hard to get, but in the end....

[birchoak]

"Raw, open, singing its heart out..." Man, you nailed it. As an English major, I keep coming back to your writing skill (which also entails an ability to sit tight and observe). That little phrase, perhaps more than anything else I've heard or written myself, best describes the Sansui 800's sound.

What most people outside of AK do not know and neglect to factor is just how much work these old receivers will need to function. To pay a professional to fully recap an 800, let alone an Eight, will likely cost more than the purchase price. And you still won't be there, because a solder joint will fail or a transistor will go bad and back in the shop it will go. You need either deep pockets & a good technician, or, a consuming passion & willingness to put in the hours of skilled labor to make a 50 year old piece roadworthy again. After 20 years of buying on Ebay, I can think of very, very few receivers, amps, cassette decks, CD players, turntables, and tuners that didn't stop working after I bought them.

Sellers state, "Works great," or, "Just serviced," but with absolutely no documentation to support the claim. We at AK know what that means, and if something has truly been gone over, it will be offered here on Barter Town or listed on Ebay for significant coin--there are no shortcuts. Once in a blue moon a lucky person (here, "lucky" means the person searches, day after day, night after night, for months and years, with an open mind, for a good deal and through sheer persistence and elapsed time on this earth, finally scores). Said lucky person then carefully researches the parts needed, summons forth the combined intellectual might of AK, and embarks upon a restoration odyssey that (best case scenario) resembles that of Stereofun's and other talented, patient AKers. Less talented, less patient buyers (me) may destroy the piece during this attempt.

To anyone fresh off a Sansui 800 Google search (not a member of AK), caveat emptor. There are no shortcuts. This is extremely difficult work for the average citizen, often with unobtanium parts, service manuals lacking key information, factory schematics with uncorrected flaws, and hours of soldering and de-soldering. You have to love it. You have to go to bed dreaming about hi-fi, and wake up still drooling over Nichicon capacitors and Bourne trimmers. Carbon resistors excite you the same way jelly beans did in your Easter basket. You may need an oscilloscope and other sophisticated tools. You should build a dim bulb tester. You should use the best solder, the best flux, and the best solder wicks on the market. You need to understand what will happen if you grab a 10,000uF/120V filter cap the wrong way, or how very dead you will be if you touch the wrong thing in a tube amplifier. You may pour all of your considerable skill into a restoration such as Stereofun's, only to discover something isn't right, and have to stay calm and figure it out. Without a powerful community like AK, I do not see anyone successfully restoring a Sanui 800, 9090DB, Eight, 2000A, 777A, etc. Not unless you are a trained technician or know one.

So, hats off to the supremely patient, technically brilliant denizens of this website who are willing to put themselves out there, on the world's biggest stage so that others may learn or simply enjoy watching top-tier talent, in the same we watched Bobby Fischer or Mark Spitz. I mean that. I think what people do here is pure magic and I have yet to tire of it.

 

[stereofun]

I will imidiatly pass on any praise for skill to the true masters of AK - have you ever followed a restoration thread from Smurfer77 or Leestereo ....and many others who are far more skilled than I. We all learn from one another, and our omnipresent English man, Hyperion, makes sure no one gets a pin-out wrong. And then we have Kevzep, the mod wizard from kiwiland, and Kale from Croatia, who once did a write up so interesting that I considered it a novel.

We are indeed blessed here at AK, and especially the Sansui forum. Cheers, Tom

P.s I think folks have had enough of the 800 lately ;-)

P.p.s English is my 2nd language. Thanks for the kind words.

 

[playittwice]

Well done, Tom.

 

[birchoak]

I will imidiatly pass on any praise for skill to the true masters of AK - have you ever followed a restoration thread from Smurfer77 or Leestereo ....and many others who are far more skilled than I. We all learn from one another, and our omnipresent English man, Hyperion, makes sure no one gets a pin-out wrong. And then we have Kevzep, the mod wizard from kiwiland, and Kale from Croatia, who once did a write up so interesting that I considered it a novel.

We are indeed blessed here at AK, and especially the Sansui forum. Cheers, Tom

P.s I think folks have had enough of the 800 lately ;-)

P.p.s English is my 2nd language. Thanks for the kind words.

Ok, I'll pipe down for a bit.

 

[smurfer77]

I will imidiatly pass on any praise for skill to the true masters of AK - have you ever followed a restoration thread from Smurfer77 or Leestereo ....and many others who are far more skilled than I. We all learn from one another, and our omnipresent English man, Hyperion, makes sure no one gets a pin-out wrong. And then we have Kevzep, the mod wizard from kiwiland, and Kale from Croatia, who once did a write up so interesting that I considered it a novel.

We are indeed blessed here at AK, and especially the Sansui forum. Cheers, Tom

P.s I think folks have had enough of the 800 lately ;-)

P.p.s English is my 2nd language. Thanks for the kind words.

I've said it before, you have a way with words that is very engaging, native English or not! I do agree we are particularly blessed her in Exclusively Sansui.

About some of the comments from Birchoak's latest post, there are some good points for sure, but I don't think we need to scare everyone off. It's not sooooo bad. But i would say one thing, an integrated amp instead of receiver has a lot less bits in it, and they are usually better spaced out and easier to work on.....

 

[birchoak]

I've said it before, you have a way with words that is very engaging, native English or not! I do agree we are particularly blessed her in Exclusively Sansui.

About some of the comments from Birchoak's latest post, there are some good points for sure, but I don't think we need to scare everyone off. It's not sooooo bad. But i would say one thing, an integrated amp instead of receiver has a lot less bits in it, and they are usually better spaced out and easier to work on.....

In an attempt to not over-discuss the Sansui 8##, have you guys heard the Sansui 2000? I don't think I've ever heard a bad thing about it and AK member Conrad has an extensive write-up on mods for it. I wonder how it compares. More $ to acquire this one; it seems to be quite popular.

I've looked at Eights for years; they are usually $350 to nearly 1K. I would have to sell at least four of my humble receivers to swing that, but if it sounds as good as people say and it's truly the zenith of build quality, it might be time to invest. As far as working on something complex, I'm a glutton for punishment.

I have flirted with the power amplifier idea for years, and an excellent Sony tuner is languishing in the closet, but I think it will be the next ice age before a 777 comes up. It sounds perfect to me: the early, low-watt receivers and amplifiers from the big names seem to be the true Golden Age of Audio, before the Wattage Wars and Specs Skirmishes (I do not listen to music with headphones on and do not need 0.0013% THD. I think that stuff looks good on paper but does not always sound good in real life). Are there other, less rare Sansui amps worthy of attention?

 

[birchoak]

I promise to leave you guys alone after this

 

[stereofun]

I have the 4000 - same deal, with factory tape still on. It is not restored but it was my first Sansui, currently spreading joy at mother in-laws house. It sounds really good, but needs a recap. I too would love to get my hands on a 2000 series and then execute the Conrad mods. That whole Bandaxal deal is interesting as it would apply to not only cap coupled units but even the Eight or the 999, the 9090, the xx1 series since those all have quasi complementary output stages, meaning all same polarity output transistors. The Eight sounds a lot like a modded 999 - to my ears these designs are a nice marriage between the later somewhat contained and well behaved "HIFI" sound versus the - here we go again - raw unfiltered cap coupled sound that unhinges itself from the mortal grip of the woofers and dances midair in front of you with a mixture of joyful purity and slight distortion.

I've had the Eight, fully restored - I now have a near pristine original condition Eight deluxe - It sounds very nice with the silkiest tweeter of them all - That too I want to restore - it is an impressive unit, but clearly subject to cost cutting compared to the 2 lbs heavier Eight.

The Eight is great sounding, but another good reason to own one is from a collectors perspective, they are so well made, so dedicated that they transcend being "a receiver" and instead becomes the expression of an era, from where it is one of the finest.

 

[birchoak]

Tom, thank you for your impressions. I hear nothing but good things about the Eight, aside from the labels coming off the buttons. That would be an epic restoration. And your descriptions of sound are a pleasure to read; you have a knack for it. I guess I shouldn't invest in an Eight, expecting the sound of the receiver that will not be mentioned here, but understand it will be different. The Yamaha CR-800 is a highly sophisticated receiver, packed to the gills with---stuff!--and it prances about with a champagne flute perfectly balanced on its upturned nose. It is a different sound, but still a very good one. More sophisticated, I guess. Maybe the Eight is like that.

 

[smurfer77]

I have had the 2000. Amazing unit. Thumping warm bass like the 350 and AU-222. Less hifi than 777 but very very good and more vintage sounding

 

[birchoak]

I will imidiatly pass on any praise for skill to the true masters of AK - have you ever followed a restoration thread from Smurfer77 or Leestereo ....and many others who are far more skilled than I. We all learn from one another, and our omnipresent English man, Hyperion, makes sure no one gets a pin-out wrong. And then we have Kevzep, the mod wizard from kiwiland, and Kale from Croatia, who once did a write up so interesting that I considered it a novel.

We are indeed blessed here at AK, and especially the Sansui forum. Cheers, Tom

P.s I think folks have had enough of the 800 lately ;-)

P.p.s English is my 2nd language. Thanks for the kind words.

Tom, I have a question regarding your restoration. I didn't know you could increase the capacitance on new replacement capacitors. I know that you can increase the voltage rating, but always heard, "Don't increase capacitance by more than 10%". Do you feel that these upgrades (Nichicon 220uF/50V replacing stock 100uF/50V cap in driver board, for example) made a big difference in the sound of the receiver? Also, are the Wima film caps a significant change in sound? I understand that this is a common upgrade for small value caps in the signal path but wondered what your personal take was (obviously you're in, but why?).

 

[maxhifi]

How do you usually deal with the component leads which have been bent flat before soldering? I melt the solder, lift the lead with a knife, let it cool a bit, then apply heat and use needle nose pliers to straighten it, then use a solder sucker. It's a bit of a challenge to do all this without damaging any traces. I wish Sansui hadn't used this method of assembly, but I presume it was all done by hand back when labor was cheaper than component insertion machines.

 

[Sansuiman]

Are you using desoldering wick, a handle operated vacuum pump/bulb, or an actual vacuum desoldering rig like a Hakko 808? I use the latter as it saves a ton of time (yes, the upfront investment is not insubstantial, but worth it) and I can usually get the tip to catch the component lead as the solder melts and use the hole in the tip to pull the lead up, and actuate the vacuum pump trigger in one motion.

Temperature setting is critical, hot enough to fully melt the solder quickly but not so much that it starts cooking the traces (or a component if it's being reinstalled later). It took me a bit of practice to get this technique down (the timing from melt to lift mostly) but now it seems be second nature.

 

[maxhifi]

Are you using desoldering wick, a handle operated vacuum pump/bulb, or an actual vacuum desoldering rig like a Hakko 808? I use the latter as it saves a ton of time (yes, the upfront investment is not insubstantial, but worth it) and I can usually get the tip to catch the component lead as the solder melts and use the hole in the tip to pull the lead up, and actuate the vacuum pump trigger in one motion.

Temperature setting is critical, hot enough to fully melt the solder quickly but not so much that it starts cooking the traces (or a component if it's being reinstalled later). It took me a bit of practice to get this technique down (the timing from melt to lift mostly) but now it seems be second nature.

Spring operated vacuum pump. I use wick in certain situations, but only when I have to because good wick costs money, and using the vac pump doesn't. I've got to get one of those Hakko 808s everyone is raving about - it really looks like the superior answer. I don't remember lifting traces recently, it's just that this kind of work is always a little touch and go - it's not so dead simple as working on a newer board with straight leads and better quality traces. Once I finally get around to it, I've got a 5000 to recap, so it's interesting to know how others do it.

 

[stereofun]

Tom, I have a question regarding your restoration. I didn't know you could increase the capacitance on new replacement capacitors. I know that you can increase the voltage rating, but always heard, "Don't increase capacitance by more than 10%". Do you feel that these upgrades (Nichicon 220uF/50V replacing stock 100uF/50V cap in driver board, for example) made a big difference in the sound of the receiver? Also, are the Wima film caps a significant change in sound? I understand that this is a common upgrade for small value caps in the signal path but wondered what your personal take was (obviously you're in, but why?).

Good questions - I cant compare the sound from 100/50 to the 220/50 - so its more about good practice. If you look at schematic, then you'll notice that C803/804 are local supply caps for the driver boards. Their job is to smooth out ripples and spikes. They are located right off the rail voltage, then to ground (these amps are single rails). When you increase their capacitance, you will likely buy yourself a bit less ripple. It's the same deal with the large reservoir cap, originally 2200. This I up-rated to 4700 for the same reason - to provide the entire amp stage a bit more stability and reserve during spikes of peak demand. Underneath the amp you'll also find a number of the axial caps that shares the same function.

Where you have to be careful are in the signal path, such as on the tone board where the uF rating might be a calculated high or low pass filter. This is where I actually had to go in and reverse a bad decision as you can read a few posts back.

Then you have coupling caps such as the C801 position on the driver board, also referred to as the input cap - these caps hands over the signal from the previous stage. Since DC is used to power the transistors so that they can amplify the AC music signal, the coupling caps job is to block residues of that DC before the next stage, and for that they utilize their innate ability as a capacitor to only pass AC. On these coupling caps, you might gain a bit more low end frequency transfer by slightly increasing the uF. For that reason I used 1.5 uF. (same deal with the output coupling caps where I went from 2200 to 3300)

Regarding Wima caps I haven't really formed an opinion by sampling before and after - to me its once more a matter of good practice to replace electrolyctics in the signal path with film whenever that's possible. The Wima MKS2 series have a nice 5mm lead spacing that makes it workable up to 10uf (in the 50VDC rating) Here you can see a lead spacing chart "PCM" - I would actually be interested in a comparison between Wima MKS and silmic II for sound, or at least other opinions on the MKS series for audio use. While polyester (mylar) from a technical standpoint, is hands down better than electrolytics, it is still not the best dielectric for carrying audio signals. I really would prefer Polypropylene, but then size becomes the issue, unless you can mount on the back or have some other tricks up your sleeve - but for the love of god, don't shoehorn one in, or allow 2 inch leads just to have a polyprop levitate mid-air
------------------------------------------

I realize the Hakko series as Sansuiman mentions is probably the right way forward. I too sit with a de-solder iron (not the spring loaded type, but the rubber pouch/balloon....or whatever that's called in English) - during the restoration I bought a new tip for it, with a smaller diameter hole and thinner walls around the circumference, this made it easier to lift off the bent legs. I also use solder wick. Tried to be careful, but still ended up with two breaches as described.

 

[birchoak]

Good questions - I cant compare the sound from 100/50 to the 220/50 - so its more about good practice. If you look at schematic, then you'll notice that C803/804 are local supply caps for the driver boards. Their job is to smooth out ripples and spikes. They are located right off the rail voltage, then to ground (these amps are single rails). When you increase their capacitance, you will likely buy yourself a bit less ripple. It's the same deal with the large reservoir cap, originally 2200. This I up-rated to 4700 for the same reason - to provide the entire amp stage a bit more stability and reserve during spikes of peak demand. Underneath the amp you'll also find a number of the axial caps that shares the same function.

Where you have to be careful are in the signal path, such as on the tone board where the uF rating are a calculated high pass or low pass filter. This is where I actually had to go in and reverse a bad decision as you can read a few posts back.

Then you have coupling such as the C801 position on the driver board, also referred to as the input cap - these caps hands over the signal from the previous stage. Since DC is used to power the transistors so that they can amplify the AC music signal, the coupling cap utilizes their innate ability as a capacitor to only pass AC, while blocking DC. On these coupling caps, you might gain a bit more low end frequency transfer by slightly increasing the uF. For that reason I used 1.5 uF. (same deal with the output coupling caps where I went from 2200 to 3300)

Regarding Wima caps I haven't really formed an opinion by sampling before and after - to me its once more a matter of good practice to replace electrolyctics in the signal path with film whenever that's possible. The Wima MKS2 series have a nice 5mm lead spacing that makes it workable up to 10uf (in the 50VDC rating) Here you can see a lead spacing chart "PCM" - I would actually be interested in a comparison between Wima MKS and silmic II for sound, or at least other opinions on the MKS series for audio use. While polyester (mylar) from a technical standpoint, are hands down better than electrolytics, it is still not the best dielectric for carrying audio signals. I really would prefer Polypropylene, but then size becomes the issue, unless you can mount on the back - but for the love of god, don't shoehorn one in, or allow 2 inch leads just to have a polyprop levitate mid-air
------------------------------------------

I realize the Hakko series as Sansuiman mentions is probably the right way forward. I too sit with a de-solder iron (not the spring loaded type, but the rubber pouch/balloon....or whatever that's called in English) - during the restoration I bought a new tip for it, with a smaller diameter hole and thinner walls around the circumference, this made it easier to lift off the bent legs. I also use solder wick. Tried to be careful, but still ended up with two breaches as described.

Tom, thank you for your detailed response---when you explain things they make sense. So, in this particular application, we can
• Increase the capacitance of the main filter cap to 4700uF without overloading anything downstream.
• Increase the capacitance of the coupling caps to 3300uF without destroying any precious transistors, etc.
• Upgrade the small value caps in the signal path to film caps; this is a no-brainer & wise men say do it.
• Keep the axial caps under the chassis the same values (this thing doesn't really have a power supply pcb, right?)
• Downgrade the 2.5A fuse to 2.0A for better protection.

Did I get it right? I like the idea of Elna Silmic IIs because Elnas seem to be the original brand of caps in a lot of this classic hi-fi; are they film caps?

 

[stereofun]

Yes - one note about the supply cap. If you were to go much higher than the 4700 uf, then you'd need to address the rectifier which could become overloaded from inrush current. in my 999 rebuild I upped the 2 supply caps (dual rail) from 4700 to 12000 - and put in a beefy rectifier.

As for the axial supply caps underneath, they are actual in various sizes - if memory serves 2 x 1000/16 v 2 x 470/35 and 1 x 330/25, 1 x 330/35 - I went all 1000uf/35 - again being lazy - those are just filter caps so the uF/Voltage rating are not super critical, unless you are the production accountant of 20,000 units and wants to save a penny when the engineers says you can. I am actually surprised Sansui would even bother to differentiate the 330/25 to the 330/35, but they have different sizes and the 35V would cost more to put in, so from an efficiency standpoint, it makes sense. Normally as you mention they would be on a separate supply board and many of the experts in here would up-rate them in a heartbeat..

I think I could be perfectly happy with Silmic II's in the signal path considering, as you say, that this was the brand Sansui used AND tweaked the sound to.

The downgraded fuse is more because that's what I had on hand....but yes, with a "new" unit that came with a blown fuse, its better to stay conservative.

 

[stereofun]

Here is the parts list - with some updates: (sorry the neat arrangement does not translate)

Sansui 800

F-1100 Driver board

Designation / Original / Notes /Replacement /Part number (D=Digikey)

C801/2 1/50 input Wima 1.5/50 505-MKS2B041501E00KO

C803/4 100/50 supply Nichi 220/50 647-UKZ1H221MHM

C805/6 100/6.3 Elna 100/16 555-RFS16V101MH3#5

C807/8 1/50 Wima 1.0/50 505-MKS2B041501E00KO

C809/10 10/50 Elna 10/50 555-RFS50V100MG3#5

C811/12 47/50 Nichi 47/50 647-UFG1H470MPM

C813/14 100/6.3 Elna 100/16 555-RFS16V101MH3#5

TR801/2 2SC281 Ksc1815 512-KSC1815YTA

TR803/4 2SC734 Ksc1815 512-KSC1815YTA

TR805/6 2SC734 driver Ksc2690 512-KSC2690AYS

TR807/8 2SA561 driver Ksa1220 512-KSA1220AYS

VR801/2 200 kOhm DC-Bal. Kept orig. (check & clean)

VR803/4 200 Ohm Bias ------” ----” -------Check lead spacing



F-1094 TONE BOARD

C705/6 1/50 Wima 1.0/50 505-MKS2B041001C00MF

C707/708 3,3/25 Wima 3.3/50 505-MKS2B043301H00KI

TR601/2 2SC458 ksc1815 512-KSC1815YTA



F-1114 POWER BOARD

C002 220/10 Nichi 330/16 647-UPW1C331MPD



F1037 EQUALIZER AMP

C601/11 1.5/15 Tantalum Wima 1.5/50 505-MKS2B041501E00KO

C603/4 33/6.3 Elna 33/6.3 555-RFS6V330ME3#5

C606/16 47/6.3 Elna 47/10 555-RFS10V470ME3#5

C607/17 10/50 Elna 10/50 555-RFS50V100MG3#5

C613/14 33/6.3 Elna 33/6.3 555-RFS6V330ME3#5

TR601/3 2SC458 Ksc1815 512-KSC1815YTA

TR602/4 2SC458 Ksc1815 512-KSC1815YTA



OTHER PARTS

C001 2200/63 Supply Nichi 4700/80 LKG1K472MESCBK‎ (D)

C003 2200/35 Tuner sup. Nichi 3300/50 LKG1H332MESBBK (D)

C821/22 2200/35 Output Nichi 3300/50 LKG1H332MESBBK (D)



AXIAL SUPPLYS UNDERNEATH

1000/16 - 470/25 - 330/35
1000/16 - 470/25 - 330/25

----- 6 x Nichicon TVX series 1000/35 ordered
from digikey 493-13377-ND

 

[birchoak]

Here is the parts list - with some updates: (sorry the neat arrangement does not translate)

Sansui 800

F-1100 Driver board

Designation / Original / Notes /Replacement /Part number (D=Digikey)

C801/2 1/50 input Wima 1.5/50 505-MKS2B041501E00KO

C803/4 100/50 supply Nichi 220/50 647-UKZ1H221MHM

C805/6 100/6.3 Elna 100/16 555-RFS16V101MH3#5

C807/8 1/50 Wima 1.0/50 505-MKS2B041501E00KO

C809/10 10/50 Elna 10/50 555-RFS50V100MG3#5

C811/12 47/50 Nichi 47/50 647-UFG1H470MPM

C813/14 100/6.3 Elna 100/16 555-RFS16V101MH3#5

TR801/2 2SC281 Ksc1815 512-KSC1815YTA

TR803/4 2SC734 Ksc1815 512-KSC1815YTA

TR805/6 2SC734 driver Ksc2690 512-KSC2690AYS

TR807/8 2SA561 driver Ksa1220 512-KSA1220AYS

VR801/2 200 kOhm DC-Bal. Kept orig. (check & clean)

VR803/4 200 Ohm Bias ------” ----” -------Check lead spacing



F-1094 TONE BOARD

C705/6 1/50 Wima 1.0/50 505-MKS2B041001C00MF

C707/708 3,3/25 Wima 3.3/50 505-MKS2B043301H00KI

TR601/2 2SC458 ksc1815 512-KSC1815YTA



F-1114 POWER BOARD

C002 220/10 Nichi 330/16 647-UPW1C331MPD



F1037 EQUALIZER AMP

C601/11 1.5/15 Tantalum Wima 1.5/50 505-MKS2B041501E00KO

C603/4 33/6.3 Elna 33/6.3 555-RFS6V330ME3#5

C606/16 47/6.3 Elna 47/10 555-RFS10V470ME3#5

C607/17 10/50 Elna 10/50 555-RFS50V100MG3#5

C613/14 33/6.3 Elna 33/6.3 555-RFS6V330ME3#5

TR601/3 2SC458 Ksc1815 512-KSC1815YTA

TR602/4 2SC458 Ksc1815 512-KSC1815YTA



OTHER PARTS

C001 2200/63 Supply Nichi 4700/80 LKG1K472MESCBK‎ (D)

C003 2200/35 Tuner sup. Nichi 3300/50 LKG1H332MESBBK (D)

C821/22 2200/35 Output Nichi 3300/50 LKG1H332MESBBK (D)



AXIAL SUPPLYS UNDERNEATH

1000/16 - 470/25 - 330/35
1000/16 - 470/25 - 330/25

----- 6 x Nichicon TVX series 1000/35 ordered
from digikey 493-13377-ND

Thank you! I had started typing up your list by hand! And I looked at some of AK member Kale's restorations...the level of cleaning and disassembly he goes to is mind numbing. I got tired just looking at the work in his threads! How he can also take the time to produce all those photographs and write it up is beyond me. I am going to follow most of your suggestions but I think I'd rather use the Elna Silmic IIs (although those red Wimas look very cool). I will not go past 4700uF capacitance for the main filter cap; indeed, I've almost always been kind of a sissy when it comes to cap values. I once upped the filter caps in a Yamaha CR-600 and destroyed it (turns out the existing, factory-installed caps were already much higher than the schematic, but I hadn't consulted AK nor the schematic first. My bad!). Compared to Kale's restorations, this one looks positively do-able.

 

[stereofun]

While enjoying the 800 very much, the tinkering goes on. I kept thinking if those S-sounds could be a bit better, there is still a bit grain, although I could live with it. Since the poly caps on the tone board helped (and perhaps since they were corrected to right value) - I went to the bench and looked at what else I had laying around. Well two Dayton polyprops 2.2uf/250 originally used on my 999 power supply caps - but removed because I could not tell a difference. Then also two brand new 1uf/250V Dayton polyprops designated "audiograde". My experience with bypassing large electro's has not been very good. I Also have a 505 with 10uf/250 attached to the two output couplers - again negligible results, perhaps a bit softer tweeter if anything, but that's not what the 505 needs. That recommendation came highly from the late great Jim "Trnsfrmr", especially for the 505 - but not to my ear.

Per username, I just like to just have fun with it - so out to the bench and fit the 2.2 on the 4700 supply

The leads can barely make it, which is good, but I need electrical tape wrapped to protect whats only a 2 mm gap between the lead and chassis.

As for the 1uF that is going unto the coupling caps its a bit more tricky - I want to keep the leads as short as possible but its rather tough to fit the rather long polyprop in between lots of wires.

Its not pretty work - and solder joints are re-flown after pic taken (I use my phone as a magnifying glass to check work)

So I come into this with negative confirmation bias, yet this make a considerable improvement, not just to the tweeter but all around. This was as unexpected as welcomed. From this point on, I'll certainly consider poly bypasses on large electro's a tweakers tool that's worth trying even if it doesn't always make a difference.