How can I pre-test my dim bulb tester?
- By canonken
- Equipment & Tools
- 23 Replies
Hi,
I just built this: https://audiokarma.org/forums/index.php?threads/dim-bulb-tester-build-and-how-to.808399/
And I think everything is fine. However, is there any real-world test (to ensure the wiring is 100% correct) I can do to pre-test it before using it to test my amp? Like plugging it into some cheap/durable electronic device and turning it on? Yes, this is a little paranoid, but more about better visualizing how it works and confirming for my education.
I did deliberately use the 'old' style switch and used the decora style outlet for...style 😀
Picture for fun, obviously this does not show the wiring!
I just built this: https://audiokarma.org/forums/index.php?threads/dim-bulb-tester-build-and-how-to.808399/
And I think everything is fine. However, is there any real-world test (to ensure the wiring is 100% correct) I can do to pre-test it before using it to test my amp? Like plugging it into some cheap/durable electronic device and turning it on? Yes, this is a little paranoid, but more about better visualizing how it works and confirming for my education.
I did deliberately use the 'old' style switch and used the decora style outlet for...style 😀
Picture for fun, obviously this does not show the wiring!
:-) :-) :-) *** Hello diyAudio members, I need some help in many questions about LM3886 *** :-) :-) :-)
Hi there ,
I'm Michalea and I'm a young girl who has chosen electronics as a hobby.😍
Please forgive my bad English, but Google Translator does its best 🙂😛
I've decided to do a bit with amplifiers, but I'm still having trouble understanding it.
Maybe you can help me to understand some things better and explain things more simply than it is described in the textbooks.
I bought a China LM3886 amplifier to do some testing and understanding.
I measured the circuit and have a few questions about it.
🤔
1) Why I need C1 at all when a cinch signal is already AC?
2) Are C1 and R2 a high pass? and how can I caculate it in a simple wise , which is the border frequenzes.
Which frequenzes makes sense to eliminate, e.g. under 60Hz ?
3) Why do the values for C1 vary between 1µF and 4.7µF in so many other designswith the same LM3886?
4) Why should C1 be made of foil and not electrolytic capacitor or ceramic, can you really hear the difference?
5) R3 and R4 is clear , the gain factor. But why do I need C2?
6) What is the R5 in only for? I've read some things that it would be better to have an RC (R8/C3) and RL (L1/R7) at the output so that the 3886 doesn't oscillate.
How do I find out the best values? ( see in circiut )
7) Then I saw that many have the supplement ( in red inside) , what is that for exactly? In the gain path?
8) Then I read some stuff about composite modes with an added OP, but I'll get to that later when I get the first things down.
9) What do you think is better for starting out as a youngster: KiCad oder EasyEda ?
10) I have a toroidal transformer with 2x 18V and 4x 6800µF and i got a light hum in Speakers, it almost sounds like the transformer
If I connect the earth of the 230V~ (gn/ge) to the GND of the circuit board, the humming becomes much quieter, but it doesn't go away.
11) If I have a good amplifier, what is the maximum V or mV that can be applied to output 3 of the LM3886 if the input is unused, i.e. open?
.... regards and from now on .....happy replying and writing 🙂
12 ) new on 29.01.2022 , phase shifting, see #14
I'm Michalea and I'm a young girl who has chosen electronics as a hobby.😍
Please forgive my bad English, but Google Translator does its best 🙂😛
I've decided to do a bit with amplifiers, but I'm still having trouble understanding it.
Maybe you can help me to understand some things better and explain things more simply than it is described in the textbooks.
I bought a China LM3886 amplifier to do some testing and understanding.
I measured the circuit and have a few questions about it.
🤔
1) Why I need C1 at all when a cinch signal is already AC?
2) Are C1 and R2 a high pass? and how can I caculate it in a simple wise , which is the border frequenzes.
Which frequenzes makes sense to eliminate, e.g. under 60Hz ?
3) Why do the values for C1 vary between 1µF and 4.7µF in so many other designswith the same LM3886?
4) Why should C1 be made of foil and not electrolytic capacitor or ceramic, can you really hear the difference?
5) R3 and R4 is clear , the gain factor. But why do I need C2?
6) What is the R5 in only for? I've read some things that it would be better to have an RC (R8/C3) and RL (L1/R7) at the output so that the 3886 doesn't oscillate.
How do I find out the best values? ( see in circiut )
7) Then I saw that many have the supplement ( in red inside) , what is that for exactly? In the gain path?
8) Then I read some stuff about composite modes with an added OP, but I'll get to that later when I get the first things down.
9) What do you think is better for starting out as a youngster: KiCad oder EasyEda ?
10) I have a toroidal transformer with 2x 18V and 4x 6800µF and i got a light hum in Speakers, it almost sounds like the transformer
If I connect the earth of the 230V~ (gn/ge) to the GND of the circuit board, the humming becomes much quieter, but it doesn't go away.
11) If I have a good amplifier, what is the maximum V or mV that can be applied to output 3 of the LM3886 if the input is unused, i.e. open?
.... regards and from now on .....happy replying and writing 🙂
12 ) new on 29.01.2022 , phase shifting, see #14
Question on Icepower 500A power connector wiring
Hi, building my first DIY amp. Planning for a 3 channel setup using a donor Parasound 5125 chassis & control boards, 500ASP + 2 ea 500A. I have the PS donor and the 500ASP in hand as well as all the harnesses. I'm waiting on the 500A's from PE to show up and I'm trying to get ahead with the wiring and prep work.
The 500ASP wiring is straight forward enough as far as I can tell. But I have a question on the power input for the 500A. Since I don't have the boards in hand I can't visually inspect or test for myself. The 500A datasheet and google fu seem to indicate that it only needs the 80V from the ASP, but looking at the connector spec, it says VP on pins 3,4 and PGND for 7,8.
So maybe I'm just overthinking this, but I'm a little worried about just trusting the documentation 100%. So according to the connector it says pins 3 and 4 get Vp, so which of these is correct:
1. The datasheet is whacked out and the hanger really should get 80V and 50V. In that case which pin is which?
2. The 500A does only need 80V and you should Y the 80V out from the ASP to both pins 3 and 4 on the 500A?
3. Pins 3 and 4 are connected on the 500A board and you only need 80V to one of them and can daisy chain the other hanger from the other pin?
4. I'm a dummy and am totally off base here.
On a side note, I found a guy with a very very cheap 5125 due to it having multiple failed channels. I was going to order a chassis for the build which was going to be about $150 plus I would need all the other parts. I got the 5125 for $100 shipped and after doing some troubleshooting I was able to save the control board and input controls, so I've been able to salvage the main power controls including the 12V trigger circuit and the auto on with signal. So coupling that with all the RCA connectors, output terminals and the nice spacious case I'm pretty pleased.
Appreciate any help with the question above.
Mike
The 500ASP wiring is straight forward enough as far as I can tell. But I have a question on the power input for the 500A. Since I don't have the boards in hand I can't visually inspect or test for myself. The 500A datasheet and google fu seem to indicate that it only needs the 80V from the ASP, but looking at the connector spec, it says VP on pins 3,4 and PGND for 7,8.
So maybe I'm just overthinking this, but I'm a little worried about just trusting the documentation 100%. So according to the connector it says pins 3 and 4 get Vp, so which of these is correct:
1. The datasheet is whacked out and the hanger really should get 80V and 50V. In that case which pin is which?
2. The 500A does only need 80V and you should Y the 80V out from the ASP to both pins 3 and 4 on the 500A?
3. Pins 3 and 4 are connected on the 500A board and you only need 80V to one of them and can daisy chain the other hanger from the other pin?
4. I'm a dummy and am totally off base here.
On a side note, I found a guy with a very very cheap 5125 due to it having multiple failed channels. I was going to order a chassis for the build which was going to be about $150 plus I would need all the other parts. I got the 5125 for $100 shipped and after doing some troubleshooting I was able to save the control board and input controls, so I've been able to salvage the main power controls including the 12V trigger circuit and the auto on with signal. So coupling that with all the RCA connectors, output terminals and the nice spacious case I'm pretty pleased.
Appreciate any help with the question above.
Mike
Texas Instruments ADC evaluation module PCM4222EVM
SOLD $114 shipped to you in a USPS Medium Flat Rate Box (CONUS only).SOLD
I bought this a few years back but never got around to a build. Brand new. Never used. Search on the gearspace.com forum as well as this forum for numerous discussions on how to DIY a state of the art ADC.
Users Guide here: https://www.ti.com/lit/ug/sbau124/sbau124.pdf?ts=1648138611241&ref_url=https%3A%2F%2Fwww.google.com%2F
“The PCM4222EVM evaluation module includes analog input and digital output circuitry with common audio connectors, providing a direct interface to audio test systems for measurement and evaluation.”
Mouser link: https://www.mouser.com/ProductDetail/Texas-Instruments/PCM4222EVM?qs=x9Fa6eo0USkoieE6GGZWGA==
Attachments
2-Way Speaker Crossover Re-Design - Arcus TM 55 - Beginners Report
TL;DR: I recently found a pair of old 2-way Speakers, made by german manufacturer Arcus. As I first listended to them, I thought that they had beautiful bass and highs, but a recessed midrange. I tried to simulate the speakers using speaker simulation software boxsim ( http://www.boxsim.de/ ) and altered the crossover, starting with old parts from the scrap bin. After a few tries i ended up with decent speakers tailored to my listening taste.
Long Report - Intro
Since i was a youth i owned a pair of old, shabby looking Arcus TM 65 3-way speakers, that i considered to sound good at that time. Two decades later i already build a few DIY audio projects, like Salas UFS Phono Preamp and Randys Aleph 30. I felt that my old 3-way speakers where too shouty for my taste. I also want to tinker with active crossovers and bi-amping later on, that’s why i was looking for a pair of 2-way speakers. I found some Arcus TM 55 2-way speakers with beautiful wooden cabinets and bought them. When i first listened to them i was pleased, because the mids seemed to be less aggressive compared to my old TM 65. But soon i realised the mids where lacking completely.
Crossover Design Arcus TM55
Arcus datasheets states that the Arcus TM55 crossover frequency is 4500 Hz. Initially i did not believe that, because regarding their 3-way designs Arcus claimed crossover frequencies of 500Hz / 2000Hz with the same model of woofer and tweeter used.
I checked the crossovers on the TM 55 and i found a fairly simple circuit: one coil acting as a first order low pass filter for the woofer and one HUGE 7uF KP capacitor used as a high pass for the tweeter. Another small coil with a 4,7 ohm resistor in series is connected in parallel to the tweeter, maybe some kind of impedance correction?
So what we have is a very simple crossover, so lets check the values of the coils and simulate the speaker using boxsim for the sake of fun and learning!
Basic Measurements
I started with stripping the crossovers apart to measure the parts values using the simple GM328 Tranistor Tester. The big coil connected to the woofer measured 1,5 mH and the small coil parallel to the tweeter 0,58 mH. So the crossover design looks like this:
I searched the internet for TSP and SPL graphs for the Vifa M21 WN-02 chassis and the Seas H174 1“ dome tweeter. I used REW to measure the impedance vs. Frequency of the chassis . With Vituix CAD i converted the graphs to asci text to make it usable for boxsim. With this data I created a first, approximate model of my 2-way speakers:
First Simulation Results (Original Crossover Design)
The frequency response graph of the first simulation showed me why this speakers sounded quite recessed in the midrange. The graph shows a bass peak between 50 and 200 Hz combined with a huge dip between 300 and 3000 Hz. The highs above 4 kHz are obviously too loud compared to the rest of the frequency spectrum.
The crossover point seems to be at approximate 1500 Hz – one third of the claimed value in the datasheet. This results look quite disappointing regarding that Arcus was a big name in the German Hi-Fi market in the 1980s.
Next Steps
Long Report - Intro
Since i was a youth i owned a pair of old, shabby looking Arcus TM 65 3-way speakers, that i considered to sound good at that time. Two decades later i already build a few DIY audio projects, like Salas UFS Phono Preamp and Randys Aleph 30. I felt that my old 3-way speakers where too shouty for my taste. I also want to tinker with active crossovers and bi-amping later on, that’s why i was looking for a pair of 2-way speakers. I found some Arcus TM 55 2-way speakers with beautiful wooden cabinets and bought them. When i first listened to them i was pleased, because the mids seemed to be less aggressive compared to my old TM 65. But soon i realised the mids where lacking completely.
Crossover Design Arcus TM55
Arcus datasheets states that the Arcus TM55 crossover frequency is 4500 Hz. Initially i did not believe that, because regarding their 3-way designs Arcus claimed crossover frequencies of 500Hz / 2000Hz with the same model of woofer and tweeter used.
I checked the crossovers on the TM 55 and i found a fairly simple circuit: one coil acting as a first order low pass filter for the woofer and one HUGE 7uF KP capacitor used as a high pass for the tweeter. Another small coil with a 4,7 ohm resistor in series is connected in parallel to the tweeter, maybe some kind of impedance correction?
So what we have is a very simple crossover, so lets check the values of the coils and simulate the speaker using boxsim for the sake of fun and learning!
Basic Measurements
I started with stripping the crossovers apart to measure the parts values using the simple GM328 Tranistor Tester. The big coil connected to the woofer measured 1,5 mH and the small coil parallel to the tweeter 0,58 mH. So the crossover design looks like this:
I searched the internet for TSP and SPL graphs for the Vifa M21 WN-02 chassis and the Seas H174 1“ dome tweeter. I used REW to measure the impedance vs. Frequency of the chassis . With Vituix CAD i converted the graphs to asci text to make it usable for boxsim. With this data I created a first, approximate model of my 2-way speakers:
First Simulation Results (Original Crossover Design)
The frequency response graph of the first simulation showed me why this speakers sounded quite recessed in the midrange. The graph shows a bass peak between 50 and 200 Hz combined with a huge dip between 300 and 3000 Hz. The highs above 4 kHz are obviously too loud compared to the rest of the frequency spectrum.
The crossover point seems to be at approximate 1500 Hz – one third of the claimed value in the datasheet. This results look quite disappointing regarding that Arcus was a big name in the German Hi-Fi market in the 1980s.
Next Steps
- simulate alternate crossover designs using boxsim
- buy a measurement microphone to measure SPL vs frequency and response of the speaker in total
- use some old stuff to do a first try in crossover design
- build it and see if something changes
- learn and have fun
Philips CD608 digital output broken
- By emanue
- Digital Source
- 3 Replies
Hi, i have an old Philips CD608 that works perfectly except digital output.
Connected to DAC, that output is completely mute. Analog output instead is working fine.
Looking at service manual, this is the interesting part:
SPDIF connector is on the lower side (BU2 -3 and BU2-4).
Service manual helps giving indication for troubleshhoting, as follow:
So i checked 3650 fuse resistor and it's ok. VDD is 4.9 V directly on PCF3523P pin 16 and 8.
Circuit 6568 is a BC848 transistor, It's not open and not shorted.
For testing i removed 3659 resistor to bypass any kind of Kill mute signal, but nothing changed with transistor base opened.
Bridge 3881, 3880 and 3889 are ok and not interrupted.
Resistor 3653 is 679ohm, so it's ok
Any other suggestion?
Are there any chances that the PCF3523P is not the culprit?
Thank you
Emanuele
Connected to DAC, that output is completely mute. Analog output instead is working fine.
Looking at service manual, this is the interesting part:
SPDIF connector is on the lower side (BU2 -3 and BU2-4).
Service manual helps giving indication for troubleshhoting, as follow:
So i checked 3650 fuse resistor and it's ok. VDD is 4.9 V directly on PCF3523P pin 16 and 8.
Circuit 6568 is a BC848 transistor, It's not open and not shorted.
For testing i removed 3659 resistor to bypass any kind of Kill mute signal, but nothing changed with transistor base opened.
Bridge 3881, 3880 and 3889 are ok and not interrupted.
Resistor 3653 is 679ohm, so it's ok
Any other suggestion?
Are there any chances that the PCF3523P is not the culprit?
Thank you
Emanuele
KEF 104 to KEF 104aB mkII with SPICE
The vintage KEF104 series loud speakers still command a lot of respect today despite its age and simplicity in design. They have been many music lovers’ favorite speakers, and will continue to be much loved.
These 70s speakers have come of age and parts are no longer supported by KEF. If the speakers have spent time in humid conditions, it would require a complete overhaul for sure, after 30years in service. The crossover needs to be checked for component accuracy. And the drivers need to be checked for the correct mechanical and electrical properties.
The crossover’s circuit board traces will have oxidation and corrosion after years of services. On close inspection and measurement, cross over tracks may have become thinner and pose a much higher series resistance in its path. In some cases, the corrosion may have introduced an open circuit.
The old black Elcap black capacitors on the cross over will have to go, even if kept from new. I have measured many of such capacitors. Even good meters usually could not determine the component’s correct values (not less, but strangely, much larger capacitance).
The old Alcap caps may still retain its value after 20 years. Modern capacitors can perform much better. There has been discussion on the series resistances of the cap that affects cross over design. There appear to be two schools of thoughts, and both have their merits.
The simple, small ferrite coils may suffer from hysteresis. At high currents, or fast changing frequencies, any ferrite core will saturate magnetically and cause a non linear departure from the ideal behavior. This will manifest as distortion. The crossovers are placed very close to the driver units (behind the metal plates). Although there could be no significant magnetic coupling path, inductors are best free from strong magnetic fields. (The selector on the KEF face plate introduce contact resistances.)
The internal wiring of these speakers is thin, and made of a few tinned cores. The fact that the crossover sits at the top right of the speaker, to facilitate selection of frequency responses, require a longer wire run. This adds to the current limiting effect of the internal wiring.
As such, the original crossover is best left as is, and a new cross over based on the old design be constructed as replacement. KEF did not put together a speaker crossover from a pile of loose parts. Individual parts must have been measured, put together, and then measured again, to meet a tight specification.
Therefore, to rebuild a crossover, we must at least have a RLC meter, SPICE software to determine the correct modern equivalent for replacement. We need to understand why the crossover was designed that way, and a measurement tool to meet that specification or goal.
Having said that, we may never be able to determine the original design goals KEF had. But KEF certainly had a design budget and there always will be a spread in specification during mass production. This is not building a pair of speakers from scratch, and we can get away from using an expensive full range SPL measuring device.
ORCAD SPICE SIMULATION
This is the original KEF 104 equivalent circuit in SPICE based on KEF’s published driver data sheets. Note that for the passive components during simulation, the inductor is always made up of an inductance and a series resistance (determined by measurements). Inductors have stray capacitance but ignored. Capacitance will have series resistance but ignored. Resistors will have series inductances, but small enough to be ignored.
As the simulation is based on the equivalent circuit, this is only an approximation of the actual voltages appearing at the driver’s terminals. This is NOT the actual SPL output of the driver as the driver has its own non linear responses. The simulation studies the electrical characteristics in the frequency domain.
The red trace shows the voltage appearing on the T27 tweeter terminals, and the green trace the voltage appearing at the B200 bass driver terminals. The blue trace shows the summed voltage on the two drivers.
The red trace shows that there is no treatment of the natural resonance at 1.1khz for the T27A driver. There is a very slight hump at the 1kHz summed output (log scale) due to this 'resonance'. This is certainly undesirable.
The green trace shows a kink at the 2.5khz region for the bass driver.
The summed output shows that from 2.5kHz to 3kHz the transition is not smooth. However, the measured SPL of the drivers may be flat (as shown by KEF) due to the non linearity of the drivers.
This is the KEF 104aB crossover.
The introduction of a 0.6uf MKT type cap at the high pass, will result in a snubbing effect at the 1.1khz region, thereby attenuates the natural resonances of the T27 driver. The T27 will start to exhibit a falling SPL above the 10khz region, so the introduction of a paralleled 10uf with 3R combo will boost the output above 10khz, and flatten the resultant measured SPL.
As modern caps have lower series resistance (low ESR), this can be compensated by a very small series R, so 3R in the high pass section can be selected to measure 3.1R or 3.2R to retain that damping effect. The value printed on the component may not be the actual value of the component.
The 0.3m inductor is realized by an air cored inductor with only 0.3R dc. The original ferrite core measured 0.55R dc. The resistance of the inductor determines how deep the attenuation of the 1.1khz notch would go. At 0R the attenuation will be close to full attenuation, ie, a short circuit. However, in a real component, there is always some resistance present. It is good to use MKT type for the 0.6uf cap that has a certain series resistance.
The choice of series inductance 2.7mH needs a low series resistance and low magnetic hysteresis. Ideally, an air cored inductor will have the least distortion as there is no core saturation. However, it is hard to find one air core inductor that gives 2.7mH and lower than 0.5R in resistance. The Mundorff Alconit or ferrite series offers a coil at 0.38R with much lower distortion at high current (high current wire).
The 0.6mH is bypassed by a 10R resistor. An air core unit with 0.1R is good choice.
The simulation of the 104aB voltages with consideration of parasitic properties of the components is as follows.
At ~1.1kHz, the voltage is snubbed and prevent the T27a from seeing the offending frequency. Due to the drivers own SPL responses, anything below 1.5khz will have very little output.
At 3kHz the knee kink as in the case of 104, is removed and the transition smoothed. This would make the speaker sound less intrusive.
The B200 transition is also smoothed near the 2-2.5kHz region. The 2.5kHz region determines the perceived sound quality of the speakers.
At 2.5khz, the summed voltages appear to be smooth and no exaggeration between the 2.5khz to above 3khz. This is a marked improvement over the 104 crossover.
The boost above 10khz for the T27 remains. The output is similar to the 104 crossover.
The analysis shows that the KEF 104aB crossover is tweaked to suit the T27a characteristics. KEF's acoustic Butterworth article might not had elaborated much in these technical details. The snubbing effect appears to be a KEF special. This same designs appears in later KEF models such as the Carlton series.
As such, it is not advisable to replace the T27a with any other driver without due consideration of the high pass circuit design. It would be better to source for a pair of used T27a (measured at 6.5ohms dc resistance) than to put in a pair of Seas or Vifa unit (resonant points can be anywhere from 900Hz to 1.2kHz, and different response curves).
Real World Components
A new cross over is constructed using direct point to point connection of components. Components are secured on a FR4 board using hot glue. The final cross over is also placed near the input terminals, far away from the magnetic field of the speaker drivers.
The main inductors can be of the Mundorff cored coil for its low series R and low hysteresis. The other two cores can be air core units but low series R as well. The inductors should be measured to ensure it is within 1% of the value stated.
The capacitors were chose to be Mundorff M caps (400-630V) types. The Mundorff caps are sonically superior to the Solen fast caps; both with a very similar cost. There is no need to bypass each cap with a 0.1 or 0.01uf cap, as this may result in a worse perceived high frequency. Individual caps should be measured for correct values.
The tolerance of the M cap is at 2%, at two extremes the values can be, in the case of 5uf theoretical, be 4.9u against 5.1u. This is not a lot, but we want to be accurate from the start. In the case of 10% tolerances, the deviation could be 4.5u or 5.5u, which will have significant shift in the cross over frequencies.
For internal wiring, it is recommended to use single core 18 or 16 gauge pure copper wires. As the crossover is nearer to the terminals, the overall wire run length is much reduced for the bass driver. The longer wire run for the T27a has little effect, as it favors a slightly higher series resistance. The coil inductance of the driver is much higher than the magnitude of the internal wiring. For the safety of the T27a, the 1 amp fast blow fuse should be in place.
The B200 drivers can still be sourced occasionally on Ebay UK. The drivers should have the same series resistance s and deviate very little from the spec of 7 ohm dc. Due to age, the driver coil may see loose particles in the air gap, or the coil former may be deformed due to heat. This can be checked by slightly depressing the cone and check for rubbing of coil on the magnets. The damaged coil may be repaired, but not many technicians are willing to do it these days.
The bass driver and the passive radiator should have the rubber suspension cleaned and treated by Armor All. The rubber suspension may become dry and brittle in some cases. The change in compliance of the suspension will surely change the sound.
The drivers must be installed with sealing gaskets/foams so that everything is air tight. Check by slightly depressing the main driver, and see a small following excursion on the bass radiator. If the bass radiator do not follow suit, there is leakage.
The new cross over would require at least 100 hours to break in, mainly due to the capacitors. New capacitors will always ‘sound’ closed in, compressed when new. Breaking in over 100 or even 200 hours will significantly improve the sound. Therefore, initial assessment of sound quality when the crossovers are brand new will surely result in a wrong conclusion.
With the new crossover installed with the good drivers, the speakers can sound close to what it was designed for, and even better.
Comparing the original 104 and new 104aB, the mid range is less bloated and high frequency is less piercing, even thought the 104 have higher series resistance caps. Bass is more extended and the speakers sound a few classes higher. The new 104aB sound more refined, fast, detailed and still musical. However, the 104 still holds its charm, mainly due to its coloration which many find it pleasing.
Comparing the original 104aB and the new 104aB, there are more details in the modern version with the equivalent sweetness. Bass is less muddled. The speakers appear to have a ‘lower’ sensitivity, as the volume control can be turned a few notches higher. In reality, the speaker can play music louder, with much lesser fatigue to the ears. The most significant different will be the speed of the speakers. The speaker will be able to handle transients much better. Cymbals and percussion sound more clear but less edgy.
As the crossovers continue to break in with the drivers, sonic spatial illusion becomes wickedly apparent. Separation of individual instruments is improved, and vocals become more life-like. Sound stage layering will be more distinct.
In conclusion, the old caps in these old speakers must be replaced. There is a choice of original Alcaps or Mundorff depending on personal preferences. The old inductors can be replaced by modern cores that will elevate the performance of the speakers. The internal wires are too thin and relocating the crossovers is advantageous (foregoing the facility of inductance change for equalization).
A new cross over is not expensive to build, and the old crossover can be retained for comparisons, or for display. This is a worthy upgrade; the returns are much greater than buying a new amplifier or even a new pair of modern speakers.
To fully appreciate the KEF 104aB, a high quality solid state amplifier of about 80-100Watts is recommended. Tube amplifiers either push pull or single ended above 10Watts can have good results as the speaker has decent efficiency. The source should be a good LP setup to fully demonstrate the capabilities of these classic speakers. KEF 104 can response up to 30kHz range. There are not many modern speakers that can sound as marvelous as these classics.
These 70s speakers have come of age and parts are no longer supported by KEF. If the speakers have spent time in humid conditions, it would require a complete overhaul for sure, after 30years in service. The crossover needs to be checked for component accuracy. And the drivers need to be checked for the correct mechanical and electrical properties.
The crossover’s circuit board traces will have oxidation and corrosion after years of services. On close inspection and measurement, cross over tracks may have become thinner and pose a much higher series resistance in its path. In some cases, the corrosion may have introduced an open circuit.
The old black Elcap black capacitors on the cross over will have to go, even if kept from new. I have measured many of such capacitors. Even good meters usually could not determine the component’s correct values (not less, but strangely, much larger capacitance).
The old Alcap caps may still retain its value after 20 years. Modern capacitors can perform much better. There has been discussion on the series resistances of the cap that affects cross over design. There appear to be two schools of thoughts, and both have their merits.
The simple, small ferrite coils may suffer from hysteresis. At high currents, or fast changing frequencies, any ferrite core will saturate magnetically and cause a non linear departure from the ideal behavior. This will manifest as distortion. The crossovers are placed very close to the driver units (behind the metal plates). Although there could be no significant magnetic coupling path, inductors are best free from strong magnetic fields. (The selector on the KEF face plate introduce contact resistances.)
The internal wiring of these speakers is thin, and made of a few tinned cores. The fact that the crossover sits at the top right of the speaker, to facilitate selection of frequency responses, require a longer wire run. This adds to the current limiting effect of the internal wiring.
As such, the original crossover is best left as is, and a new cross over based on the old design be constructed as replacement. KEF did not put together a speaker crossover from a pile of loose parts. Individual parts must have been measured, put together, and then measured again, to meet a tight specification.
Therefore, to rebuild a crossover, we must at least have a RLC meter, SPICE software to determine the correct modern equivalent for replacement. We need to understand why the crossover was designed that way, and a measurement tool to meet that specification or goal.
Having said that, we may never be able to determine the original design goals KEF had. But KEF certainly had a design budget and there always will be a spread in specification during mass production. This is not building a pair of speakers from scratch, and we can get away from using an expensive full range SPL measuring device.
ORCAD SPICE SIMULATION
This is the original KEF 104 equivalent circuit in SPICE based on KEF’s published driver data sheets. Note that for the passive components during simulation, the inductor is always made up of an inductance and a series resistance (determined by measurements). Inductors have stray capacitance but ignored. Capacitance will have series resistance but ignored. Resistors will have series inductances, but small enough to be ignored.
As the simulation is based on the equivalent circuit, this is only an approximation of the actual voltages appearing at the driver’s terminals. This is NOT the actual SPL output of the driver as the driver has its own non linear responses. The simulation studies the electrical characteristics in the frequency domain.
The red trace shows the voltage appearing on the T27 tweeter terminals, and the green trace the voltage appearing at the B200 bass driver terminals. The blue trace shows the summed voltage on the two drivers.
The red trace shows that there is no treatment of the natural resonance at 1.1khz for the T27A driver. There is a very slight hump at the 1kHz summed output (log scale) due to this 'resonance'. This is certainly undesirable.
The green trace shows a kink at the 2.5khz region for the bass driver.
The summed output shows that from 2.5kHz to 3kHz the transition is not smooth. However, the measured SPL of the drivers may be flat (as shown by KEF) due to the non linearity of the drivers.
This is the KEF 104aB crossover.
The introduction of a 0.6uf MKT type cap at the high pass, will result in a snubbing effect at the 1.1khz region, thereby attenuates the natural resonances of the T27 driver. The T27 will start to exhibit a falling SPL above the 10khz region, so the introduction of a paralleled 10uf with 3R combo will boost the output above 10khz, and flatten the resultant measured SPL.
As modern caps have lower series resistance (low ESR), this can be compensated by a very small series R, so 3R in the high pass section can be selected to measure 3.1R or 3.2R to retain that damping effect. The value printed on the component may not be the actual value of the component.
The 0.3m inductor is realized by an air cored inductor with only 0.3R dc. The original ferrite core measured 0.55R dc. The resistance of the inductor determines how deep the attenuation of the 1.1khz notch would go. At 0R the attenuation will be close to full attenuation, ie, a short circuit. However, in a real component, there is always some resistance present. It is good to use MKT type for the 0.6uf cap that has a certain series resistance.
The choice of series inductance 2.7mH needs a low series resistance and low magnetic hysteresis. Ideally, an air cored inductor will have the least distortion as there is no core saturation. However, it is hard to find one air core inductor that gives 2.7mH and lower than 0.5R in resistance. The Mundorff Alconit or ferrite series offers a coil at 0.38R with much lower distortion at high current (high current wire).
The 0.6mH is bypassed by a 10R resistor. An air core unit with 0.1R is good choice.
The simulation of the 104aB voltages with consideration of parasitic properties of the components is as follows.
At ~1.1kHz, the voltage is snubbed and prevent the T27a from seeing the offending frequency. Due to the drivers own SPL responses, anything below 1.5khz will have very little output.
At 3kHz the knee kink as in the case of 104, is removed and the transition smoothed. This would make the speaker sound less intrusive.
The B200 transition is also smoothed near the 2-2.5kHz region. The 2.5kHz region determines the perceived sound quality of the speakers.
At 2.5khz, the summed voltages appear to be smooth and no exaggeration between the 2.5khz to above 3khz. This is a marked improvement over the 104 crossover.
The boost above 10khz for the T27 remains. The output is similar to the 104 crossover.
The analysis shows that the KEF 104aB crossover is tweaked to suit the T27a characteristics. KEF's acoustic Butterworth article might not had elaborated much in these technical details. The snubbing effect appears to be a KEF special. This same designs appears in later KEF models such as the Carlton series.
As such, it is not advisable to replace the T27a with any other driver without due consideration of the high pass circuit design. It would be better to source for a pair of used T27a (measured at 6.5ohms dc resistance) than to put in a pair of Seas or Vifa unit (resonant points can be anywhere from 900Hz to 1.2kHz, and different response curves).
Real World Components
A new cross over is constructed using direct point to point connection of components. Components are secured on a FR4 board using hot glue. The final cross over is also placed near the input terminals, far away from the magnetic field of the speaker drivers.
The main inductors can be of the Mundorff cored coil for its low series R and low hysteresis. The other two cores can be air core units but low series R as well. The inductors should be measured to ensure it is within 1% of the value stated.
The capacitors were chose to be Mundorff M caps (400-630V) types. The Mundorff caps are sonically superior to the Solen fast caps; both with a very similar cost. There is no need to bypass each cap with a 0.1 or 0.01uf cap, as this may result in a worse perceived high frequency. Individual caps should be measured for correct values.
The tolerance of the M cap is at 2%, at two extremes the values can be, in the case of 5uf theoretical, be 4.9u against 5.1u. This is not a lot, but we want to be accurate from the start. In the case of 10% tolerances, the deviation could be 4.5u or 5.5u, which will have significant shift in the cross over frequencies.
For internal wiring, it is recommended to use single core 18 or 16 gauge pure copper wires. As the crossover is nearer to the terminals, the overall wire run length is much reduced for the bass driver. The longer wire run for the T27a has little effect, as it favors a slightly higher series resistance. The coil inductance of the driver is much higher than the magnitude of the internal wiring. For the safety of the T27a, the 1 amp fast blow fuse should be in place.
The B200 drivers can still be sourced occasionally on Ebay UK. The drivers should have the same series resistance s and deviate very little from the spec of 7 ohm dc. Due to age, the driver coil may see loose particles in the air gap, or the coil former may be deformed due to heat. This can be checked by slightly depressing the cone and check for rubbing of coil on the magnets. The damaged coil may be repaired, but not many technicians are willing to do it these days.
The bass driver and the passive radiator should have the rubber suspension cleaned and treated by Armor All. The rubber suspension may become dry and brittle in some cases. The change in compliance of the suspension will surely change the sound.
The drivers must be installed with sealing gaskets/foams so that everything is air tight. Check by slightly depressing the main driver, and see a small following excursion on the bass radiator. If the bass radiator do not follow suit, there is leakage.
The new cross over would require at least 100 hours to break in, mainly due to the capacitors. New capacitors will always ‘sound’ closed in, compressed when new. Breaking in over 100 or even 200 hours will significantly improve the sound. Therefore, initial assessment of sound quality when the crossovers are brand new will surely result in a wrong conclusion.
With the new crossover installed with the good drivers, the speakers can sound close to what it was designed for, and even better.
Comparing the original 104 and new 104aB, the mid range is less bloated and high frequency is less piercing, even thought the 104 have higher series resistance caps. Bass is more extended and the speakers sound a few classes higher. The new 104aB sound more refined, fast, detailed and still musical. However, the 104 still holds its charm, mainly due to its coloration which many find it pleasing.
Comparing the original 104aB and the new 104aB, there are more details in the modern version with the equivalent sweetness. Bass is less muddled. The speakers appear to have a ‘lower’ sensitivity, as the volume control can be turned a few notches higher. In reality, the speaker can play music louder, with much lesser fatigue to the ears. The most significant different will be the speed of the speakers. The speaker will be able to handle transients much better. Cymbals and percussion sound more clear but less edgy.
As the crossovers continue to break in with the drivers, sonic spatial illusion becomes wickedly apparent. Separation of individual instruments is improved, and vocals become more life-like. Sound stage layering will be more distinct.
In conclusion, the old caps in these old speakers must be replaced. There is a choice of original Alcaps or Mundorff depending on personal preferences. The old inductors can be replaced by modern cores that will elevate the performance of the speakers. The internal wires are too thin and relocating the crossovers is advantageous (foregoing the facility of inductance change for equalization).
A new cross over is not expensive to build, and the old crossover can be retained for comparisons, or for display. This is a worthy upgrade; the returns are much greater than buying a new amplifier or even a new pair of modern speakers.
To fully appreciate the KEF 104aB, a high quality solid state amplifier of about 80-100Watts is recommended. Tube amplifiers either push pull or single ended above 10Watts can have good results as the speaker has decent efficiency. The source should be a good LP setup to fully demonstrate the capabilities of these classic speakers. KEF 104 can response up to 30kHz range. There are not many modern speakers that can sound as marvelous as these classics.
TU-8500 with Mundorf Cap
Finally, I use Mcap to replace the stock PP caps
0.47uf X 4 Mcap Supreme MKP
0.1 uf X 2 EVO Silver Gold Oil
0.47uf Silver & Gold Oil is too big. It is very difficult to install in TU-8500. I use Mcap Supreme MKP instead.
0.47uf X 4 Mcap Supreme MKP
0.1 uf X 2 EVO Silver Gold Oil
0.47uf Silver & Gold Oil is too big. It is very difficult to install in TU-8500. I use Mcap Supreme MKP instead.
Extremely Unconventionally Shaped Dual 15" Sealed Sub Enclosure Build
- Car Audio
- 39 Replies
Attachments
ISO Phono stage impedance question
- By Wang1
- Analogue Source
- 4 Replies
Anyone by any chance have schematic of iso phono stage, i want to change the input impedance.
The blue resistor on the left is 100r and on the right 10r, the wide area of the light green of PCB is ground plane. So i am guessing that the blue resistor is the one setting the input impedance, if that is the case is it a simple matter of changing this resistor?
The blue resistor on the left is 100r and on the right 10r, the wide area of the light green of PCB is ground plane. So i am guessing that the blue resistor is the one setting the input impedance, if that is the case is it a simple matter of changing this resistor?
Reverseengineering an HK_Ottala
- Solid State
- 8 Replies
Power amps are something very close to my music heart.
I have quite a lot of them around here, Zen in different versions, Accuphase, own designs etc, and my favourite for the latest few years, an ML11, the last power amp Tom Collangelo designed, a magical little amp that can drive anything and equiped with quite a lot of power despite its modest sepcs.
As all Mark Levinson amps it has a certain sonic signature, leaning towards the darker kind of voice, never bright but with brutal transientresponse and attack when delivered such.
A while ago, rather long while actually, i traded a pair of JBL HF drivers and horns for an interesting amp, as i if i needed more of these.
Utterly sophisticated design, B&O has something to learn here,.
But to the reason i wanted it. This amp is designed by Matti Ottala during his time as chief engineer at Harman Kardon.
The information i´ve got says it was designed as a prototype for the Citation XX project and the same story/rumour says 3 or 5 of this amp was built at HK workshop.
The amp is among the best i´ve ever heard with an extreme lightness combined with unbelievable power and transientresponse , a "sharp" voice but still not bright, drives anything i´ve loaded it with the same ease.
One major flaw, as an extremely wide bandwith amp it is very prone to amplify and distribute to speakers any wire or airborne signals in the vicinity.
Fabulous buildquality.
So, i will dismante this amp, primarily for replacing wiring and also check/replace caps and find the right place to insert lp filtering to
avoid interference.
If there is interest in recreating this fantastic amp, it actually is just that, among the best i´ve ever listened to, ill start a thread reengineering the amp and se what we can end up with.
I have quite a lot of them around here, Zen in different versions, Accuphase, own designs etc, and my favourite for the latest few years, an ML11, the last power amp Tom Collangelo designed, a magical little amp that can drive anything and equiped with quite a lot of power despite its modest sepcs.
As all Mark Levinson amps it has a certain sonic signature, leaning towards the darker kind of voice, never bright but with brutal transientresponse and attack when delivered such.
A while ago, rather long while actually, i traded a pair of JBL HF drivers and horns for an interesting amp, as i if i needed more of these.
Utterly sophisticated design, B&O has something to learn here,.
But to the reason i wanted it. This amp is designed by Matti Ottala during his time as chief engineer at Harman Kardon.
The information i´ve got says it was designed as a prototype for the Citation XX project and the same story/rumour says 3 or 5 of this amp was built at HK workshop.
The amp is among the best i´ve ever heard with an extreme lightness combined with unbelievable power and transientresponse , a "sharp" voice but still not bright, drives anything i´ve loaded it with the same ease.
One major flaw, as an extremely wide bandwith amp it is very prone to amplify and distribute to speakers any wire or airborne signals in the vicinity.
Fabulous buildquality.
So, i will dismante this amp, primarily for replacing wiring and also check/replace caps and find the right place to insert lp filtering to
avoid interference.
If there is interest in recreating this fantastic amp, it actually is just that, among the best i´ve ever listened to, ill start a thread reengineering the amp and se what we can end up with.
some "noob" questions (frequency response/excursion vs. group delay)
- Subwoofers
- 17 Replies
Hello 🙂
I am playing arround with hornresp for some time now, trying to find the best enclosure for my DD Audio 608D2.
When i kinda understand now how to get what i want... i just realised... i do not really know WHAT i want.
So i got a few questions, you guys hopefully can answer me.
I got a pretty huge car (4.1m x 1.9m x 1.5m inside), in wich i want to use this sub in. So kinda low resonance and not that stiff chassis.
Lets start with the following image, how do i want it to look in my car if i want a mostly flat in car response? Do i want the sub to play up to 200Hz as seen in the images? Or is it better to crossover earlier?
So, more precise questions to this:
Do i want more like a curve at the end, like in "D", or more like a slope like in "A" ?
Do i want the upper end above 200Hz more like "C" going down continiously, or do i want it like in "B" with a as big as possible steep gap?
Also another thing i realised while simulating, since i am tuning below fs(38Hz), i get a lot of excursion OR group delay.
Is it allways low group delay vs high excursion / high group delay vs low excursion? Or is there anything i could do, do keep excurion AND group delay low?
Oh, one more question... my driver has a one direction xmax of 13mm and a xmech 32mm ... what would be a save value (hornresp displacement) to use it at low frequencys like 30Hz and below? (simulating at my amps rms output of 600w)
20mm? 25mm? Or should i defently stay at xmax?
Thanks in advance
Benni
I am playing arround with hornresp for some time now, trying to find the best enclosure for my DD Audio 608D2.
When i kinda understand now how to get what i want... i just realised... i do not really know WHAT i want.
So i got a few questions, you guys hopefully can answer me.
I got a pretty huge car (4.1m x 1.9m x 1.5m inside), in wich i want to use this sub in. So kinda low resonance and not that stiff chassis.
Lets start with the following image, how do i want it to look in my car if i want a mostly flat in car response? Do i want the sub to play up to 200Hz as seen in the images? Or is it better to crossover earlier?
So, more precise questions to this:
Do i want more like a curve at the end, like in "D", or more like a slope like in "A" ?
Do i want the upper end above 200Hz more like "C" going down continiously, or do i want it like in "B" with a as big as possible steep gap?
Also another thing i realised while simulating, since i am tuning below fs(38Hz), i get a lot of excursion OR group delay.
Is it allways low group delay vs high excursion / high group delay vs low excursion? Or is there anything i could do, do keep excurion AND group delay low?
Oh, one more question... my driver has a one direction xmax of 13mm and a xmech 32mm ... what would be a save value (hornresp displacement) to use it at low frequencys like 30Hz and below? (simulating at my amps rms output of 600w)
20mm? 25mm? Or should i defently stay at xmax?
Thanks in advance
Benni
Separate subwoofer volume control
- By 0bserver
- Subwoofers
- 3 Replies
Separate volume control for sub amp?
Help please, this might be obvious and I may be overthinking it. I finished my Swan speaker build (loving them). Currently my setup is on my PC. I want to use the main volume on the Focusrite for the mains and the headphone volume control as a dedicated volume control for the Subwoofers.
USB > Focusrite 2i2 > Cleanbox Pro > Dayton SA1000 ( Bass AMP) > Nad 326bee (Main Amp)
The Focusrite has a separate volume control for the 1/4" dedicated headphone jack. The signal is full range though. What piece of gear do I need to strip down the full range 1/4" signal to an LFE or subwoofer amp signal (RCA preferably)?
WTB Exicon ECF20N25A w/ green colour code (>2.8Ids)
Would anyone have a few pcs they could part with? Used ones are perfectly fine.
Pretty please with sugar lumps on top?
Pretty please with sugar lumps on top?
Tamradio Tamura Sony TC-500A Recycled Transformers
- By Simply Audio
- Tubes / Valves
- 19 Replies
Hi,After a couple of Sony TC-500A tear downs I have recycled the Tamradio Tamura Nickel Output transformers and Sony Holey Basket FE103A Alnico speakers .
The Configuration consists of a 6J5 driver tube and Single Ended EL84 output tube running in Pentode mode 5 watts output.
The Cathode of the EL84 is set to 40mA with a LM317 regulator and a 22v Zenner Diode to set the Control Grid Voltage .
A EZ81 rectifier tube using a low noise Toroidal Power Transformer .
Power Supply Capacitors are all Film
Other parts are Dale Wire Wound , Kiwame Carbon , Caddock thick film Resistors and Russian K40Y signal capacitors .
Hook up wire is mainly Kimber Teflon OFC Copper .
Input RCA and Speaker Terminals KLE Low Mass Copper .
The Sony Holey Baskets- Fostex FE103A / Alnico speakers are housed in a copy of the Humble HiFi Solo 103 cabinets, http://www.humblehomemadehifi.com/download/Humble Homemade Hifi_Solo-103.pdf
The System has excellent resolution and a very natural sound, Vocals have a real lifelike quality as do musical instruments , probably due to the Tamradio Vintage Permalloy Output transformers and high quality parts.
I prefer this amp to the Audion 300B for naturalness of human voice also pictured here .
Lower bass frequencies can't be expected without a sub woofer .
The Chassis is from Hammond as are the DC chokes which are dual rail past the first two smoothing capacitors .
In summary a great sounding second system from recycled parts .
Attachments
Problems on Chinese Power amplifer Module
- By Mongis
- Solid State
- 31 Replies
Hello All
I have a problem on a Chinese amplifier which uses 2SC5200 and 2SA1943 transistors the final transistors were purchased from Farnell and are of the same category of HFE (o) the amplifier works but the loudspeaker safety trips as soon as the we push the volume I try to cool the input transistors but nothing changes. Does the problem come from the too spartan loudspeaker security or is it possible that the amplifier has a breakdown? I wired 2 modules the 2 do the same thing. Thanks for your help.
I have a problem on a Chinese amplifier which uses 2SC5200 and 2SA1943 transistors the final transistors were purchased from Farnell and are of the same category of HFE (o) the amplifier works but the loudspeaker safety trips as soon as the we push the volume I try to cool the input transistors but nothing changes. Does the problem come from the too spartan loudspeaker security or is it possible that the amplifier has a breakdown? I wired 2 modules the 2 do the same thing. Thanks for your help.
VK Music will display at High End Show in Munich
I will demo in High End Show in Munich - Halle 4 T-08
I am looking forward to meet you all soon
I am looking forward
I am looking forward to meet you all soon
I am looking forward
Stubbed horn questions / usable bandwith, phaseshift, crossover
- Subwoofers
- 2 Replies
Hello guys,
when simulating different enclosure types with my driver, i endet with a stubbed horn now.
I got the Space for it in my car (roughly about 200l), and since it is a van, i guess my roomgain to start relatively low. My main goal was to have a as flat as possible frequency response down to ~20Hz.
Since i have no experience at all in building subwoofers or any other hifi components, i got some questions:
Thanks for the help 🙂
Benni
when simulating different enclosure types with my driver, i endet with a stubbed horn now.
I got the Space for it in my car (roughly about 200l), and since it is a van, i guess my roomgain to start relatively low. My main goal was to have a as flat as possible frequency response down to ~20Hz.
Since i have no experience at all in building subwoofers or any other hifi components, i got some questions:
- First... will this be a usable subwoofer? What do you think about it?
- could i use it WHITOUT any high/lowpass filters, just the way it is now? (Excursion is ok at my amps max power)
- does it make sense, to use a subwoofer up to like 200Hz, like it would be in this case?
- how would the crossover for it work? At wich frequency should i crossover, and how would that work with the delays/phaseshifts in the upper frequency?
Thanks for the help 🙂
Benni
For Sale Precision Matched Harris / Fairchild SFP9610 MOSFETs
- By rhthatcher
- Swap Meet
- 0 Replies
