Jump to content
The Klipsch Audio Community
Sign in to follow this  
Jeffrey D. Medwin

KT88 Direct-Coupled Design

Recommended Posts

      4-26-19 

 

Made up a pair of 9.52 feet ( two Bob Fulton 57 1/8th inch lengths ) of SINGLE Mil Spec 8 AWG speaker leads , and inserted them into the VOTT A7-8 system. Used two leads, a lead for each polarity, so I was listening to a  " speaker-wire mis-matched " stereo, due only to speaker wire differences ( left VS right channel ). 

 

I compared the Single 8 AWG  on one speaker, to the pseudo- 8 AWG leads, consisting of a Trio of 12+12+14 AWG Mil Spec., on the other speaker.. 

 

Well, the new Single 8 AWG Mil Spec / Copper-Silver wire probably out-performs the " Trio " 8 AWG . Needs 100 hours break - in, to really tell. Highs are not yet fully-developed.

 

This means that the recent idea, ( to re-do the KT88 amp's main ground wire, 3/4s of an inch out of the Power Transformer winding, using a Single 8 AWG length ), is likely going to be a very good thing to do.

 

Next to do is the wiring ( amp to both loudspeakers ) 100% with Single 8 AWG Mil Spec., right up to the Voice Coil's Terminals of each ALTEC 515B fifteen inch woofer. 

 

1757410949_P1010090editeddate.thumb.jpg.9133af6897bfa20ff98fe6b325b9a3b6.jpg

 

 

 

This photo above is the last section of wiring to install.  Its a 57 1/8th inch pair of m22759/11 8 ( 8 AWG ) Mil Spec wire.  This pair goes from the input of my right speaker's crossover ( the copper lugs bolt to the crossover's Magnum C5237-4 Terminal Block )  and the opposite bare / tinned end, gets soldered DIRECTLY to the two terminals that  hold the woofer ( ALTEC 515B ) voice coil wires.  A direct wire path, 8 AWG, amplifier  to 515B driver.   This is substantial wire, 133 strands of copper 29 AWG, each strand silver plated ( for wide bandwidth and conductivity ) , teflon jacketed, all made to military standards.   Then .there is 100 hours to play music through it, so it gets broken - in.

 

798956389_MagnumSNIPPED2.thumb.jpg.f10d7113ecd7ec44a53248f7a021f865.jpg

 

 

                                       Brass studs, also use brass washers and brass buts, with Magnum C5237-4 in many places in the system.

 

 

There is NO crossover being used on the speakers' 515Bs, ......I operate both of the 15's full range.  From 135 hZ. on up, the 15 inch driver is front-horn loaded and direct-radiating......and I just love hearing that !! ..  Then, the ALTEC 802D High frequency driver is " layed - in " ,..... above each 15 inch driver. 

 

What does this wire have to do with the KT88 amp build??   Its important that the wiring be excellent.    Almost all amp-building determinations get decided by ear, and the wiring is part of the equation and system.  My first audio mentor, Mr. Robert Fulton, used to tell me on many occassions " an inch of bad wire can ruin the musical experience." 

 

I am so happy to have the new Weller D650 soldering gun my Mentor Dennis told me to get.  It can do 300 Watts, and I am learning how to use it better now.  Its mandatory to own and use ... to do this type of build !! 

 

Jeffrey

Share this post


Link to post
Share on other sites

                                   4-28-19  Progress Photos

 

On the KT88 amp prototype grounding buss wire,  today -  I had time to remove / replace  the previous 8 AWG Mil Spec "TRIO "  ( 12+12+14 ) and installed a SINGLE run of Mil Spec 8 AWG wire.  

 

The  new single 8 AWG wire starts an inch or so out of the Power Transformer's High Voltage winding's Center Tap .

 

The VOTT A7-8 speakers played-back reference music for over ten hours, using the new Single 8 AWG wire.   I KNOW its an improvement over ANY " trio " 12+12+14  wiring that was used before.  Its a learning process.  I learned previously (  from my audio mentor's input,  and combining experimentally  12+12+14 AWG  in parallel ),  that 8 AWG was a nice amount of wire - for optimized performance.  Now,  .......the last two days, I have learned........ its best done as a single run of 8 AWG Mil Spec. 

 

Still need 90 hours or so more time for break-in of the speaker wire, to make a final decision on the single wire's high end's playback.

 

The High Efficiency speakers LOVE high efficiency ( meaning this type of 8 AWG Mil Spec ) wiring. 

 

The professional-quality Altec drivers respond dynamically in a lovely manner, but only when all connections ( amp and crossover, crossover out ) are torqued down TIGHT with a socket wrench, can one seem to get it all, dynamically - contrast - wise.

 

 

 

 

P1010102.thumb.JPG.57d727a843cf37e19b520ec5da87b91a.JPG

           

    12 AWG  lead-out ( short white ) is a remnant of the original center tap lead.  The to-be-added ( a red ) 8 AWG lead out -  it is to be this KT88 amp's                                                                                                                               main ground bus.

 

 

 

 

 

 

 1823524161_P1010010nedited4.thumb.jpg.e593ee8a8fc31967d87938db74eb7982.jpg

 

                                        The shortened white ( center tap )12 AWG Mil Spec wire had absorbed solder, ( from the prior 12+14 being grafted onto it.  )

                                         It was un-bendable, stiff, so this was not perfection,  but rather, the best soldering that could be done  The Weller D650 ,

                                                                            on the 300 Watt trigger-setting, made for a nice shiney solder joint.

 

 

 

 

 

                                                          933573805_P1010028edited2.thumb.jpg.f7532c58017457d05cf667d9d30bd2ed.jpg

 

 

     What we see ( above ) is likely ...... the first tube amp ever being built with a wide-band, low-loss  8 AWG Mil Spec ( copper/silver ) ground buss,.........

     starting from out of the Power Transformer's High Voltage Center Tap. 

 

     This possibly also is  the first directly-coupled KT88 amp ever built.   'Hard to tell.

 

     Certainly, this is the first directly-coupled KT88 amp with my audio mentor,  Dennis Fraker's  / Serious Stereo L.S.E.S.power supply in it.            

 

  

 

 

P1010006.thumb.JPG.0c55c15ce263c5d1772ad011f0648941.JPG

                                                                                                                       

                                                                                                       

                                                                                            Much more to do !!   Thanks - for taking your time - to view this.

 

Share this post


Link to post
Share on other sites

Other than the output impedance issue, distortion will be ridiculously high and gain will be off the charts high.  You'll get that sound of the AA-5 table radio though, if that's what you're after.

 

Feedback is an acceptable compromise when running a single ended pentode amp.  While you are welcome to hate it all you want, that doesn't somehow negate the plethora of reasons that you need to use it in an amp of this type.  Being an "expert" in the field, I can't imagine how you don't already know this?

Share this post


Link to post
Share on other sites

If we take your driver mu of 100 and reduce it due to loading (call it 65), then add in the KT-88 pentode loaded, you have a power amp with a gain of 65dB (assuming a resistive load).

 

If your GPA-604 has an impedance peak or two at low frequencies (we'll say 50 ohms for the sake of illustration), then gain will be 82dB where that impedance peak is.  This will not be remedied by short wiring runs, a steel chassis, or a proprietary power supply.  If you want these results, I'd spend $100 on a car audio amplifier and put a battery on the floor in your listening room, then flip the "bass boost" switch.

 

This of course assumes that you adequately size the cathode bypass cap, which isn't a given all things considered.  If you go too small, by say believing that the cathode bypass cap can be sized by using the first order speaker crossover filter calculation, then things can get worse!

Share this post


Link to post
Share on other sites
3 hours ago, audiowize said:

If we take your driver mu of 100 and reduce it due to loading (call it 65), then add in the KT-88 pentode loaded, you have a power amp with a gain of 65dB (assuming a resistive load).

 

If your GPA-604 has an impedance peak or two at low frequencies (we'll say 50 ohms for the sake of illustration), then gain will be 82dB where that impedance peak is.  This will not be remedied by short wiring runs, a steel chassis, or a proprietary power supply.  If you want these results, I'd spend $100 on a car audio amplifier and put a battery on the floor in your listening room, then flip the "bass boost" switch.

 

This of course assumes that you adequately size the cathode bypass cap, which isn't a given all things considered.  If you go too small, by say believing that the cathode bypass cap can be sized by using the first order speaker crossover filter calculation, then things can get worse!

 

Hello Hello !!

 

Thanks for posting.   If you read this thread, you will discover I use a VOTT A7-8 with a 515B and a 802D, ...... not a GPA 604 . 

 

The KT88 needs to do about 1/12th to 3/4s of a Watt, and distortion will be low !!   I use 1/4 of a Watt 6005 amplifier now, and can " almost"  ( but not really ) get by.

 

I am familiar with these ordinary EE concepts you are discussing.   I have already thought about all you bring up, discussed it thoroughly with smart people who have decades of practical experience,  ( eg: removing feedback loops from many many professionally-designed pentode tube amps ) .  Any and all you bring up, such as the wisdom of using a high mu driver, etc etc. is known to me.  

 

Yet, I am building the KT88 DC amp the way I want to, ......despite ( not in -spite ) of what " common knowledge " you just posted.  I am not worried one iota.   

 

It is true, I will have to hear this, to find out what it does.  No one has built this, this way , before.   That  - is what makes this build interesting to me !!

 

I will be able to report honestly as to what the final result is. 

 

If I like this effort, and use it, you will always be welcomed to visit and bring any amp you like over for an A-B amp-comparing session.   We can accurately be the final judge.... only by judging what we hear. 

 

I say, its not a matter at all of what you, or anyone else,  believes to be theoretically true

 

If you feel confident, continue to build and listen to your favorite designs in tube amps .   Be my guest.  I'll not protest.

 

What works, works.  Its just that simple. 

 

Jeffrey

Share this post


Link to post
Share on other sites
11 minutes ago, Jeffrey D. Medwin said:

 

I say, its not a matter at all of what you, or anyone else,  believes to be theoretically true.  " The proof of the pudding. "

 

Theory is proven and measurable.  A builder building your zero feedback pentode amp on the dark side of the moon will run into the same problems that I would building the amp here.  That is the distinction, it's not a belief at all!  Again, I will ask how you can be an "expert" in tube amplification and not understand how a pentode works?  This is a contradiction.  

 

On the other hand, a lot of what you posted here is a flat earth argument.  You believe many things that are demonstrably false.  Your conviction is so strong that you don't care how many forums you get kicked off of, you'll just keep going!  There are people who knock on my door with strange books who also hold this same conviction.  

 

 

Share this post


Link to post
Share on other sites
1 hour ago, audiowize said:

 

Theory is proven and measurable.  A builder building your zero feedback pentode amp on the dark side of the moon will run into the same problems that I would building the amp here.  That is the distinction, it's not a belief at all!  Again, I will ask how you can be an "expert" in tube amplification and not understand how a pentode works?  This is a contradiction.  

 

On the other hand, a lot of what you posted here is a flat earth argument.  You believe many things that are demonstrably false.  Your conviction is so strong that you don't care how many forums you get kicked off of, you'll just keep going!  There are people who knock on my door with strange books who also hold this same conviction.  

 

 

 

Sorry, 

 

Listen - Up :

 

The person guiding me has several decades of practical experience removing negative feedback loops form Professional-use pentode amps, with positive results, in real-world venues. 

 

I am not here to argue with you. 

 

I certainly never ever stated anywhere  I am an " expert" in tube amplification, did I??                

 

This design does go beyond the ordinary, by the book, understanding.  Relax.  Lets get it built and evaluate it !!    Will be fun. 

 

I have already properly addressed prior Forums ( twice here ) , and changes made, .................you maybe missed that also.   

 

Chill-out some,  don't be hostile.  Enjoy the day !! 

 

Jeffrey

Share this post


Link to post
Share on other sites
1 hour ago, Jeffrey D. Medwin said:

The KT88 needs to do about 1/12th to 3/4s of a Watt, and distortion will be low !!   I use 1/4 of a Watt 6005 amplifier now, and can " almost"  ( but not really ) get by.

 

 

 

This is the key to why fb is not necessary in Jeff''s particular application.  At such a small a fraction of the amp's maximum power output, distortion will be low as stated.  Of course, as the power output increases, so will the distortion.  This is certainly measurable.  Given the level at which many forum members listen, the power demands will be such that distortion will be excessively high.  It would be easy enough to add some voltage fb at the output tube to deal with that issue.

 

Maynard

Share this post


Link to post
Share on other sites

" It would be easy enough to add some voltage fb at the output tube to deal with that issue." 

 

100% agree Maynard

 

I would add strapping the output pentode as triode to control the output impedance.

 

Dan

Share this post


Link to post
Share on other sites

It is also worth mentioning that the silicone you are using to glue stuff down is corrosive and will oxidize copper rapidly.  

Share this post


Link to post
Share on other sites
18 hours ago, tube fanatic said:

 

This is the key to why fb is not necessary in Jeff''s particular application.  At such a small a fraction of the amp's maximum power output, distortion will be low as stated.  Of course, as the power output increases, so will the distortion.  This is certainly measurable.  Given the level at which many forum members listen, the power demands will be such that distortion will be excessively high.  It would be easy enough to add some voltage fb at the output tube to deal with that issue.

 

Maynard

This says nothing of the damping issues with doing this.  Even if Jeff makes 3/4 of a watt at 1kHz, the impedance hump at low frequencies will cause the amp to put out far more power than this.  If we presume that the amp will deliver that 0.75W of power into an 8 ohm load, and the load pops up to 50 ohms around the driver resonance, now the amp will try to make 20+W (edit, will rise significantly in terms of voltage).  In reality the impedance peaks are probably much higher than 50 ohms, but it's sufficient to explain why this is such a terrible idea. Your only prayer is that the DS-025 will limit how out of control this gets, and the primary inductance will put a cap on how nuts the output gets at low frequencies.  On the high end, and amp like this is the kind of amp that should never be run without a speaker connected.

 

You also have a lot of confidence in the amount of power Jeff thinks he can get away with.  A 98dB driver will struggle with less than a watt unless your listening level is very, very low.  With the topology chosen for this amp, you would also want to make sure there's no LF content! 

 

Share this post


Link to post
Share on other sites
10 hours ago, audiowize said:

This says nothing of the damping issues with doing this.  Even if Jeff makes 3/4 of a watt at 1kHz, the impedance hump at low frequencies will cause the amp to put out far more power than this.  If we presume that the amp will deliver that 0.75W of power into an 8 ohm load, and the load pops up to 50 ohms around the driver resonance, now the amp will try to make 20+W.  In reality the impedance peaks are probably much higher than 50 ohms, but it's sufficient to explain why this is such a terrible idea. Your only prayer is that the DS-025 will limit how out of control this gets, and the primary inductance will put a cap on how nuts the output gets at low frequencies.  On the high end, and amp like this is the kind of amp that should never be run without a speaker connected.

 

You also have a lot of confidence in the amount of power Jeff thinks he can get away with.  A 98dB driver will struggle with less than a watt unless your listening level is very, very low.  With the topology chosen for this amp, you would also want to make sure there's no LF content! 

 

 

I don’t want to bog down Jeff’s thread with too much technical discussion which we could do until the cows come home.  Power output is unlikely to exceed around 40% of the quiescent dissipation, and a very high impedance reflected back from the speaker will reduce output power.  We can get into discussing the output transformers later on if Jeff says he won’t mind.  

 

Maynard

Share this post


Link to post
Share on other sites
2 hours ago, tube fanatic said:

 very high impedance reflected back from the speaker will reduce output power. 

 

Clipping is still clipping.  If you're output voltage rises to 20V because the load impedance moves up, this is still clipping behavior.  

Share this post


Link to post
Share on other sites

Here's some of the basic math that governs these relationships.  You and Jeff are welcome to take the position that your willful ignorance is just as valid as 60 years of engineering, bIut for everyone else on here or any future readers, it's important to keep in mind. 

 

The relationship mu=gm*Rp still holds true for a pentode, but the Rp is the plate impedance of the pentode in parallel with the loading impedance.  We will use the KT-88 data, the DS-025 data, and the Klipsch RF-7 II data to do some analysis on what may happen.  Jeff is notorious for dismissing any opinions on what he's up to because we don't have his exact speakers, but this is the Klipsch forum right?

 

The RF-7 III is an 8 ohm nominal speaker with a small dip at 4 ohms and a few peaks around 50 ohms.  The DS-025 is a 2.5K:8 series feed output transformer, and the KT88 is a pentode with a gm of about 0.011mA/V and an Rp of about 12K.

 

At 8 ohms, the DS-025 will reflect back 2.5K, and in parallel with Rp that is about 2060 ohms. The amplification factor of the output stage will be 22.6x under these conditions.

At 4 ohms, the DS-025 will reflect back 1.25K and in parlallel with Rp that is about 1132 ohms.  The amplification factor of the output stage will be 12.5x under these conditions.

At 50 ohms, the DS-025 will reflect back 15.6K, and in parallel with Rp that is about 6872 ohms.  The amplification factor of the output stage will be 75.6x under these conditions.

 

Next we consider that for all of these wanderings of the output, the input signal stays the same. Since you mentioned power, here is the power (and the impedance in question) assuming the amp is putting out 1W at 8 ohms.

 

At 8 ohms, 1W 

At 4 ohms, 0.6W

At 50 ohms, 1.8W

 

You have admitted that the amp is only good for 2.83V at the output, but at 50 ohms we are up around 9.5V of output.  This is way, way beyond the "1W" operating level. From your own admission, the amp is only good for 0.16W at those impedance peaks.

 

So again, how does power level go down when load impedance goes up? 

Share this post


Link to post
Share on other sites
On 4/30/2019 at 3:15 PM, tube fanatic said:

 

This is the key to why fb is not necessary in Jeff''s particular application.  At such a small a fraction of the amp's maximum power output, distortion will be low as stated.  Of course, as the power output increases, so will the distortion.  This is certainly measurable.  Given the level at which many forum members listen, the power demands will be such that distortion will be excessively high.  It would be easy enough to add some voltage fb at the output tube to deal with that issue.

 

Maynard

 

 

Hi Maynard, 

 

 I think you have a pretty good handle on where I am going with this design.   How refreshing, especially since you have the " conventional " EE background.   Kudos to you.

 

I want to point out again, that this is an amp I am designing for me and for my use, on my speaker load, to please ( or not ) me in the end.   I am simply showing any interested folks my build in this thread, because it likely will contain some new ideas that they can possibly try on their own, possibly to their listening benefit.

 

This thread is NOT at all meant as a wholesale amp suggestion for use on Klipsch speakers.   

 

The typical Klipsch speaker will often have a more complex crossover ( harder to drive !! ) and furthermore, it may often be a three-way crossover.  My VOTT A7-8 speaker load, as presently configured, is far simpler.  No crossover what-so-ever on the 515B 15 inch woofer, and second order filter on the 802D High Frequency driver........... thats all there is !! 

 

In my estimation, this amp would  work " OK to wonderful " on a well-wired Klipsch speaker, IF I simply made a single ( very easy ) change - connected the KT88 as a psuedo triode finals stage. 

 

Recall please, everyone, , from my earlier post, I intend to build this amp first as a Pentode,  do long-term listening evaluations, then possibly / maybe listen to it pseudo triode connected for the KT88.   

 

What is lost or sorely-underestimated to some people is...... this amp will be unlike all other pentodes in existence !!

 

It will have a LSES supply, and a separate Shunt Regulated ... and ....  Final - Filtered ( L/C ) G2 pentode screen supply .  And it only will operate at about 1/12th to 3/4s of a Watt ..........on my speaker load.  It will have no " out-of-time " negative feedback loop, (  an  E.E.'s cure, that is worse than the sickness. )  I just have to build it, and listen to it critically, to find out how it plays back music. 

 

Until its built, no one knows what it will sound like.   Its an audio experiment.  My audio mentor,  ( who happens to build great 2A3 amps ), is highly encouraging me to execute this design, and feels it has great potential to please.  We shall find out.

 

I hope this post makes the design goals I follow herein crystal-clear, to avoid continued confusion by any of the traditional " by the book " thinkers. 

 

Maynard, I appreciate your contributions  a whole lot !!  Thank you.  This particular KT88 amp will be fun to build and to hear. 

 

Jeffrey

 

 

Share this post


Link to post
Share on other sites

                   5-1-19  Progress Photos

Got as bit more done, made up AC cord, and anchored it into chassis.  I use Chris Ven Haus's ( VH Audio)  https://www.venhaus1.com/diymains.html     FLAVOR 2 Power Cord, as a general idea. and pay more attention to other parts of the design which I feel are much more important, to overall music playback performance.

 

                                                          1640714029_P1010035editeduseFIRST.thumb.jpg.dc509f423f88e2e7be7eae0265530d71.jpg

 

                                                                            Some of the Parts I chose to use.  10 AWG Mil Spec is also a good wire choice for this.

 

 

                                          1456647491_SnippedPicture2.JPG.ef4e8f9c1b621cdd6e1a8aa5619293b0.JPG

 

                                                                     Heavy Duty Rubber Grommet , plus I like the AC cord in the plastic tubing, as additional protection.

 

 

   1809117730_P1010045use6th.thumb.jpg.d36a47c7eaa66adc6cf36f5c0c2e680e.jpg

 

                                                

                 

 

                     1177530774_SnippedPicture4.thumb.JPG.ea45a4d857fbfc1d8365110c7b6bfc50.JPG

   

Closest switch is Power ON-OFF.  Rear - most AC switch is illegal in a commercial product.  It bypasses the amp's fuse and CL-90 sort-start turn-on thermistor, for a more "direct" connection to my home's power grid.  I call this my " Attended Listening Switch ", and I only have it on WHEN I AM IN THE ROOM, LISTENING . 

 

A DIYer can do this, at one's own riskIt is totally illegal in a commercial product, as it bypasses all safety codes.  But OHH, those improved drum transients, on Steve Clarke's " Solo Drums " CD, when I switch it in !!  A LSES supply is a source-feed supply, and I like having this.  Plus, supply caps exist  in my amps that can do 1.500 Amperes Instantaneous peak-wise.  Sonically, such caps are enjoyed, uniquely heard, as they define the very leading edge of all sharp music transients.

 

 

 

 

                                                                                       8001201_5-1-19ToDoEdited5.jpg.936be73d09eba99fd1ea7eed8c2214ce.jpg

 

                                                                                                               Highlighted above, to wire next ! 

 

 

 

Share this post


Link to post
Share on other sites
On 4/30/2019 at 8:08 PM, audiowize said:

It is also worth mentioning that the silicone you are using to glue stuff down is corrosive and will oxidize copper rapidly.  

 

 

" Worth mentioning " ?????? 

 

"  80 grit sandpaper and a palm sander gets the finish on the brass plate.    Clear coated afterwards, so it won't tarnish over the years.  "

 

Thank you.

Share this post


Link to post
Share on other sites

If you're wiring your amplifier with copper wire, the outgassing of that silicone bathroom caulk will cause rapid oxidation. 

Share this post


Link to post
Share on other sites
On 4/8/2019 at 7:03 AM, CECAA850 said:

The proper way to create a good solder joint is to make a solid mechanical connection first

You my dear friend would be shocked then to see how most modern electronic devices are made. Solder "paste' is applied with a silkscreen type device to the circuit board. Then a robotic machine picks up and places the devices precisely into position so the legs (and lots of devices have no legs just little balls) are sitting in the paste. Then the board goes through a "reflow" oven which melts the paste into its components flux and solder. In effect there is no "mechanical" connection other than the intermetallic bond of the solder between the electrodes on the part and the circuit board pads.

Share this post


Link to post
Share on other sites

How much open loop gain will this amp have? And at what frequency?

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.
Sign in to follow this  

×
×
  • Create New...