You're probably correct. My HE1000 Stealth headphones get up to 93dB on a single mW. And my HD6X0 series Sennheiser headphones get >100dB on a single mW. Typical outputs from contemporary headphone amps, though, far exceed what would be needed to power them. And I'm not really sure why this is. Is it a marketing thing, where consumers chase bigger numbers for no good reason? I don't know. But look at an entry-level tube hybrid like the Schiit Vali 3, for example. It's $150 and puts 1.5W into 32Ohms and 400mW into 300 Ohms. That's pretty typical for the market these days. Their higher-end tube-hybrid, the Lyr, gives 6W into 32Ohms and 900mW into 300 Ohms, and goes for $600.
I should just buy a Lyr. Idk. That's boring compared to what I'd been thinking about. But hell, maybe I'll let you talk me out of this.
What do you think? Why do modern headphone amps drive so much power?
Okay, so now you have all the direction of the AI, I have one question you may want to put to it/him/her.... Where do I start? Let Ai tell you to or where to start the design and then kick it off. No one can tell you where that is a decision on your own. Until AI learns to imagine, it can only be interrogated on things that are. It is only a big encyclopedia at the moment. Ask the AI what it would sound like for instance or what it/his/her preference in music is, what brand of capacitor will be most suitable for the input. Why di it noty suggest an ECC83 for the tube based on its sound quality.
Anyway, it is time to start your journey..... into the abyss. 😱
Anyway, it is time to start your journey..... into the abyss. 😱
And what do you know? AI is fact based, it has no perception it only deals with what is. 🤣🤣🤣
Some of you seem to be getting kind of emotional. I hope we can keep the conversation to the point.
(What do you think? Why do modern headphone amps drive so much power?) Heck knows, maybe they were guided by AI. Interestingly, you are starting to doubt AI's reasoning. Or maybe it was what you asked "I want an amp that can drive 1W into 32 Ohms." You could have asked it for a 1000 watt into 32 Ohms and it would simply comply. Again everything else will follow, it would suggest a power supply of 150VAC at 2 amps. etc. etc.
Okay so let us start with your power requirements for your headphones, your level of listening and the genre(s) you like.
Why not have multiple taps or switches for optimum performance in both cases? Depending on how well optimized you want it, you could end up building quite different amplifiers in one box to suit your specific 32 ohm cans, as well as your specific 300 ohm ones.
You could build an all-in-one design, but if you're unaware of the limitations, you could have a lot of redundant current drain for the 300 ohm headphones, as well as excessively high voltage rails for the 32 ohm ones.
It sounds like that youtube video where that one guy used a special trick to do the thing!
I listen to a wide variety of music, from orchestral classics to jazz to modern rock and hip-hop. My main headphones are my HE1000 Steatlh (93dB/mW, 32Ohms) planar magnetics. I also use HD6XX, 600, and 660S2 headphones, which yield between 102 and 104dB/mw into 300Ohms nominal), though not at much. I just would prefer to have an amp that can drive them so that I don't feel the need to get rid of them. I bought them over the course of several years, before recently making the jump into hi-fi with the HEK St.
I have an SPL meter, and put the mic into the open cup of the headphone to adjust the volume so that the average with pink noise is about 70-75dB. I have two main use cases: critical listening (HEK St), usually no more than 2-3 hours at a time, but several times a week, if not daily, and quiet listening in the background while working on other things (the Senns usually are less distracting for this use), and that can be all-day, every day, sometimes.
This thread has helped me become more acutely aware of the limitations of OpenAI's current models. I have not have much time with the o1-preview, however, as I'm limited to 30 prompts every 2 weeks with it. I should regain access to that model tomorrow. It has not so far been involved in this project, because I used it for other things. I have mostly used the GPT specialized for electrical hardware engineering / design, which is based on 4o. I have also sometimes used o1-mini-preview to think through the specialized GPT's suggestions. I do see, though, that while the calculations are reliable, and the general knowledge of the components is good enough to make specific recommendations for whatever is needed in that part of the circuit, the ability of the AI to understand the overall design is spotty. Sometimes the context window is large enough for it to grasp much of what's going on. But it's not reliable, at all. And the more the context window gets filled in, the spottier the reasoning ability.
I do believe these issues will be resolved with ChatGPT 5, or whatever their next level reasoning / agentic AI will be called. But it isn't ready, yet. And while I knew I was in over my head, I didn't know my life preserver had lead weights randomly attached at different times. I just didn't know enough about what the AI was telling me to see that it was forgetting one piece of the design while working on some other piece, meaning that some things were left out entirely, and some things were mismatched with the rest of the design. While it's disappointing, it has been a great learning experience for me so far.
I have an SPL meter, and put the mic into the open cup of the headphone to adjust the volume so that the average with pink noise is about 70-75dB. I have two main use cases: critical listening (HEK St), usually no more than 2-3 hours at a time, but several times a week, if not daily, and quiet listening in the background while working on other things (the Senns usually are less distracting for this use), and that can be all-day, every day, sometimes.
This thread has helped me become more acutely aware of the limitations of OpenAI's current models. I have not have much time with the o1-preview, however, as I'm limited to 30 prompts every 2 weeks with it. I should regain access to that model tomorrow. It has not so far been involved in this project, because I used it for other things. I have mostly used the GPT specialized for electrical hardware engineering / design, which is based on 4o. I have also sometimes used o1-mini-preview to think through the specialized GPT's suggestions. I do see, though, that while the calculations are reliable, and the general knowledge of the components is good enough to make specific recommendations for whatever is needed in that part of the circuit, the ability of the AI to understand the overall design is spotty. Sometimes the context window is large enough for it to grasp much of what's going on. But it's not reliable, at all. And the more the context window gets filled in, the spottier the reasoning ability.
I do believe these issues will be resolved with ChatGPT 5, or whatever their next level reasoning / agentic AI will be called. But it isn't ready, yet. And while I knew I was in over my head, I didn't know my life preserver had lead weights randomly attached at different times. I just didn't know enough about what the AI was telling me to see that it was forgetting one piece of the design while working on some other piece, meaning that some things were left out entirely, and some things were mismatched with the rest of the design. While it's disappointing, it has been a great learning experience for me so far.
Yeah. I may end up focusing solely on the 32Ohms for my HEK St, and just using my MCTH or Fiio K7 to drive the Sennheisers. I really don't want to compromise on the quality of what's feeding my HEK St.
I could just buy a Schiit Vali 3. It meets my requirements into both kinds of headphones. But it's a $150 amp. It was designed to make money. I want a no-compromises amp, using only the best parts. For example, the thing that started all this was that my MCTH's volume pot sucks audibly. My first thought was just to replace it with a nice ALPS. But then I'm afraid of burning it, or ruining the entire amp somehow. I'd rather keep what I've got, and try to add to my amp collection with something that's as good for being an amplifier as the HEK St is for being a headphone. But maybe the Schiit Vali 3 is already better than anything I could make. Idk. It's certainly a whole lot cheaper. I don't want its compromises, money-saving measures.
So can we say that your worse case is 93 dB/mW (in other words the HE1000 is least loud?) Can you turn up the volume until your eyes water and measure the SPL.
The HE1000 Stealth is the least sensitive, but the differences in resistance mean that it and the Senns come out sounding very near each other in volume, if I unplug one and plug in the other without adjusting the volume knob. Let me go ahead and do a quick, rough experiment. I'll put on the HEKSt with some pink noise, and adjust the volume on the MCTH to average around 75dB. Then, I'll plug in the HD6XX and take the SPL reading.
HEK St: ~71dB
HD6XX: `73dB
That's probably well within the margin of error for this quick and dirty approximation of an experiment. Subjectively, the difference is too small for me to be sure I'm hearing at all, as I have to unplug one set, plug in the other, and change headphones on my head. By that time, my ability to compare is pretty much shot, when the results are this close.
HEK St: ~71dB
HD6XX: `73dB
That's probably well within the margin of error for this quick and dirty approximation of an experiment. Subjectively, the difference is too small for me to be sure I'm hearing at all, as I have to unplug one set, plug in the other, and change headphones on my head. By that time, my ability to compare is pretty much shot, when the results are this close.
Last edited:
Putting on Sting's "All This Time" on the HEK St, I turned it up to louder than I'd ever want to spend much time hearing it. It measured peaks around 83dB in fast A-weighted. I doubt I'm playing transients much louder than 93dB, which means 1mW should be enough to power the HEK St.
And yet, my MCTH can drive over 1W into them. Amir at ASR measured 1.33W into a 32 Ohm load from the MCTH.
And yet, my MCTH can drive over 1W into them. Amir at ASR measured 1.33W into a 32 Ohm load from the MCTH.
I think the time has come to publish some schematics.
The one in your first post does not make sense 🙂
The one in your first post does not make sense 🙂
POWER LOW VOLTAGE SUPPLY
Okay, let us say that you want one watt power into a 30 ohm load. Equates to +-6VDC + MOSFET drop = 15V. You want a split rail so let us say +15 - 0 -15V. I think that LM7815 and LM7915 would be more than sufficient for power supply regulation. You want about 3V drop across the regulators so +-18V from the diode bridge. There will be 1,4V drop across the diode pairs which means the transformer secondaries will be safe to say 15VAC each. A popular value. You need 1 watt per channel and the amp is 50% efficient thus you need 2 x 1 watt x 150% = 3 watt loosely translates to a 5VA transformer.
Rounded up, per rail, I will use 1000uF for the reservoir cap followed by 100 uF after the regulator and some HF caps of say 100 nF across each electrolytic.
This was worked out in my head while typing, so confer with AI, I might be off the mark completely. You can do the hi voltage side similarly.
EDIT: You can ask AI to work these values perhaps much better than my guesstimates, build it test it an confirm it works, put it aside and get onto the amp.
EDIT2: I am not designing the power supply, just showing how to go about it, there is no paraphernalia like CRC or CLC hanging off it anywhere, no lights switches, arc suppression anything. Just a stab in the dark, detail is yours my friend.
Okay, let us say that you want one watt power into a 30 ohm load. Equates to +-6VDC + MOSFET drop = 15V. You want a split rail so let us say +15 - 0 -15V. I think that LM7815 and LM7915 would be more than sufficient for power supply regulation. You want about 3V drop across the regulators so +-18V from the diode bridge. There will be 1,4V drop across the diode pairs which means the transformer secondaries will be safe to say 15VAC each. A popular value. You need 1 watt per channel and the amp is 50% efficient thus you need 2 x 1 watt x 150% = 3 watt loosely translates to a 5VA transformer.
Rounded up, per rail, I will use 1000uF for the reservoir cap followed by 100 uF after the regulator and some HF caps of say 100 nF across each electrolytic.
This was worked out in my head while typing, so confer with AI, I might be off the mark completely. You can do the hi voltage side similarly.
EDIT: You can ask AI to work these values perhaps much better than my guesstimates, build it test it an confirm it works, put it aside and get onto the amp.
EDIT2: I am not designing the power supply, just showing how to go about it, there is no paraphernalia like CRC or CLC hanging off it anywhere, no lights switches, arc suppression anything. Just a stab in the dark, detail is yours my friend.
Last edited:
Thank you so much! Here's what I did. I opened a new chat with o1-mini-preview. I fed it the article you linked as audioexpress about the Zenkido amp. I told it we were adapting that design to power HEK St headphones. Then, I gave it your text. Here is what it returned:
You're on the right track with your initial power supply design for adapting the ZenKido amplifier to drive your HE1000 Stealth headphones. However, there are a few areas where refining your calculations and component selections can enhance performance, ensure reliability, and maintain safety. Let's walk through a detailed, step-by-step guide to optimize your low-voltage power supply design.
---
## 1. Clarifying Power Requirements
### a. Understanding the Load: HE1000 Stealth Headphones
### b. Desired Output Power
You mentioned aiming for 1 watt per channel into a 30Ω load. However, for your HE1000 Stealth headphones (32Ω), this translates to:
\[
P = V^2 / R \implies V = \sqrt{P \times R} = \sqrt{1 \times 32} \approx 5.66V_{\text{RMS}}
\]
But considering your goal of a tubey, rich sound with pristine quality, ensuring sufficient headroom is essential. Therefore, aiming for 1W per channel is reasonable.
### c. Total Power Consideration
Assuming an efficiency of 50% (typical for linear amplifiers), the input power (P_in) required is:
\[
P_{\text{in}} = \frac{P_{\text{out}}}{\text{Efficiency}} = \frac{2W}{0.5} = 4W
\]
Adding a safety margin of 150% to accommodate transient spikes and ensure transformer longevity:
\[
\text{Transformer VA} = 4W \times 1.5 = 6VA
\]
Recommendation: Opt for a 10VA transformer to provide ample headroom and ensure stable operation.
---
## 2. Transformer Selection
### a. Voltage Rating
You're targeting a ±15V dual-rail supply. After full-wave bridge rectification, the DC voltage (\( V_{\text{DC}} \)) can be approximated as:
\[
V_{\text{DC}} = V_{\text{RMS}} \times 1.414 - 2 \times V_{\text{diode drop}}
\]
Assuming a 15VAC secondary:
\[
V_{\text{DC}} = 15V \times 1.414 - 1.4V \approx 19.8V
\]
### b. VA Rating
As calculated:
\[
\text{Required VA} \geq 6VA \quad \text{(Recommended: 10VA)}
\]
### c. Example Transformer
Note: Ensure the transformer is center-tapped to facilitate the creation of the split-rail supply.
---
## 3. Rectification and Filtering
### a. Rectifier Configuration
- Full-Wave Bridge Rectifier: Utilizes four diodes to convert AC to DC.
Component Suggestion:
- Bridge Rectifier IC: KBPC5010 or equivalent, capable of handling the required current and voltage.
### b. Voltage Drops
### c. Reservoir Capacitors
To smooth the rectified DC:
- Per Rail: 1000µF electrolytic capacitor.
Recommendation:
- Voltage Rating: At least 25V (preferably 35V for safety).
### d. Post-Regulation Filtering
After the linear regulators:
Placement:
- Close to Regulators: To effectively filter out high-frequency noise.
---
## 4. Voltage Regulation
### a. Selecting Regulators
### b. Voltage Drop Across Regulators
### c. Power Dissipation in Regulators
\[
P_{\text{reg}} = V_{\text{drop}} \times I_{\text{out}}
\]
Assuming 1A per rail:
\[
P_{\text{reg}} = 4.8V \times 1A = 4.8W
\]
Recommendation:
---
## 5. Capacitor Selection and Sizing
### a. Reservoir Capacitors (Before Regulators)
### b. Post-Regulator Capacitors
### c. High-Frequency Decoupling
### d. Calculating Filter Capacitor Size
To ensure minimal ripple voltage, the capacitor size should be sufficient based on the load current and the desired ripple.
\[
\Delta V = \frac{I}{f \times C}
\]
Where:
Example Calculation:
Assuming:
\[
\Delta V = \frac{1}{120 \times 1000 \times 10^{-6}} \approx 8.3V
\]
This high ripple is unacceptable. Therefore, increasing capacitor size or reducing load current is necessary.
Adjusted Calculation for Acceptable Ripple (~1V):
\[
C = \frac{I}{f \times \Delta V} = \frac{1}{120 \times 1} \approx 8333µF
\]
Recommendation:
- Increase Reservoir Capacitors: Use 10,000µF per rail to reduce ripple effectively.
---
## 6. Transformer VA Rating Correction
Your initial estimation suggested a 5VA transformer, but based on detailed calculations, a 10VA transformer is more appropriate. This ensures:
---
## 7. Practical Implementation Steps
### a. Components List
1. Transformer:
- Dual 15VAC center-tapped, 10VA
2. Rectification:
- Bridge Rectifier (e.g., KBPC5010)
- Four 1N4007 diodes (if building your own bridge)
3. Filtering:
- Electrolytic Capacitors:
- 10,000µF, 25V per rail
- Post-Regulator:
- 100µF, 25V per rail
- Decoupling:
- 100nF ceramic capacitors per rail
4. Voltage Regulators:
- LM7815 for +15V
- LM7915 for -15V
- Heatsinks for both regulators
5. Additional Components:
- Fuse: 1A slow-blow for input protection
- NTC Thermistor: For inrush current limiting
- PCB or Breadboard: For assembly
- Wire and Connectors: As needed
### b. Assembly Steps
1. Transformer Setup:
- Connect the primary to mains (ensure correct and safe wiring).
- Identify the center-tap as the ground (0V).
2. Bridge Rectifier:
- Connect the secondary windings to the bridge rectifier.
- Ensure correct polarity: AC inputs to ~, + and - outputs appropriately.
3. Reservoir Filtering:
- Connect 10,000µF capacitors to each rail post-rectification.
- Observe correct polarity to prevent capacitor damage.
4. Voltage Regulation:
- Connect LM7815 to the +V rail.
- Connect LM7915 to the -V rail.
- Attach heatsinks to both regulators.
- Add 100µF capacitors post-regulation for stability.
5. Decoupling:
- Place 100nF capacitors across each electrolytic capacitor pair near the regulator inputs and outputs.
6. Protection:
- Insert a 1A slow-blow fuse on the input side to protect against overcurrent.
- Add an NTC thermistor in series with the transformer primary for inrush current limiting.
7. Testing:
- Before Powering Up:
- Double-check all connections.
- Verify no shorts exist between rails.
- Power Up:
- Use a variac or a current-limited power supply for initial testing.
- Measure DC voltages to ensure correct ±15V outputs.
- Check for excessive heat on regulators.
---
## 8. Fine-Tuning and Validation
### a. Ripple Voltage Measurement
Use an oscilloscope to check the ripple voltage on the output rails. Aim for:
- ±15V Rail: Ripple < 1V RMS
If ripple is higher, consider increasing the reservoir capacitance or improving filtering.
### b. Load Testing
### c. Audio Performance Testing
---
## 9. Safety Considerations
---
## 10. Summary of Adjusted Component Values and Recommendations
1. Transformer:
- Voltage: Dual 15VAC center-tapped
- VA Rating: 10VA (Recommended: 10VA or higher)
2. Rectification:
- Bridge Rectifier: KBPC5010 or equivalent
- Diodes (if separate): Four 1N4007
3. Filtering:
- Reservoir Capacitors: 10,000µF, 25V per rail
- Post-Regulator Capacitors: 100µF, 25V per rail
- Decoupling Capacitors: 100nF ceramic per rail
4. Voltage Regulators:
- Positive Rail: LM7815 with heatsink
- Negative Rail: LM7915 with heatsink
5. Protection:
- Fuse: 1A slow-blow
- NTC Thermistor: For inrush current limiting
6. Additional:
- High-Frequency Caps: 100nF ceramics across electrolytics
- Heat Sinks: Appropriately sized for LM7815 and LM7915
---
## 11. Additional Tips
---
## 12. Final Validation and Testing
After assembling the power supply:
1. Initial Power-On:
- Use a variac or current-limited source to gently apply power.
- Verify DC voltages before connecting to the amplifier.
2. Load Testing:
- Connect a resistor (e.g., 15Ω) to simulate the headphone load.
- Monitor voltage stability and temperature.
3. Audio Testing:
- Connect the headphones.
- Play audio at various volume levels.
- Listen for clarity, absence of hiss or hum, and overall sound quality.
4. Iterative Refinement:
- Based on testing, adjust capacitor sizes or regulator configurations if necessary.
- Ensure all components operate within safe thermal and electrical limits.
---
By following this refined approach, you can create a robust, low-voltage power supply tailored for your headphone amplifier needs. This ensures that your HE1000 Stealth headphones receive clean, stable power, preserving their pristine and tubey sound characteristics while maintaining low total harmonic distortion and noise.
Feel free to reach out if you need further assistance or clarification on any of these steps!
You're on the right track with your initial power supply design for adapting the ZenKido amplifier to drive your HE1000 Stealth headphones. However, there are a few areas where refining your calculations and component selections can enhance performance, ensure reliability, and maintain safety. Let's walk through a detailed, step-by-step guide to optimize your low-voltage power supply design.
---
## 1. Clarifying Power Requirements
### a. Understanding the Load: HE1000 Stealth Headphones
- Impedance (R_load): 32Ω
- Sensitivity: 93dB/mW
### b. Desired Output Power
You mentioned aiming for 1 watt per channel into a 30Ω load. However, for your HE1000 Stealth headphones (32Ω), this translates to:
\[
P = V^2 / R \implies V = \sqrt{P \times R} = \sqrt{1 \times 32} \approx 5.66V_{\text{RMS}}
\]
But considering your goal of a tubey, rich sound with pristine quality, ensuring sufficient headroom is essential. Therefore, aiming for 1W per channel is reasonable.
### c. Total Power Consideration
- Per Channel: 1W
- Channels: 2 (Stereo)
- Total Output Power (P_out): 2W
Assuming an efficiency of 50% (typical for linear amplifiers), the input power (P_in) required is:
\[
P_{\text{in}} = \frac{P_{\text{out}}}{\text{Efficiency}} = \frac{2W}{0.5} = 4W
\]
Adding a safety margin of 150% to accommodate transient spikes and ensure transformer longevity:
\[
\text{Transformer VA} = 4W \times 1.5 = 6VA
\]
Recommendation: Opt for a 10VA transformer to provide ample headroom and ensure stable operation.
---
## 2. Transformer Selection
### a. Voltage Rating
You're targeting a ±15V dual-rail supply. After full-wave bridge rectification, the DC voltage (\( V_{\text{DC}} \)) can be approximated as:
\[
V_{\text{DC}} = V_{\text{RMS}} \times 1.414 - 2 \times V_{\text{diode drop}}
\]
Assuming a 15VAC secondary:
\[
V_{\text{DC}} = 15V \times 1.414 - 1.4V \approx 19.8V
\]
### b. VA Rating
As calculated:
\[
\text{Required VA} \geq 6VA \quad \text{(Recommended: 10VA)}
\]
### c. Example Transformer
- Primary: 120V (or your local mains voltage)
- Secondary: Dual 15VAC center-tapped (for ±15V)
- VA Rating: 10VA
Note: Ensure the transformer is center-tapped to facilitate the creation of the split-rail supply.
---
## 3. Rectification and Filtering
### a. Rectifier Configuration
- Full-Wave Bridge Rectifier: Utilizes four diodes to convert AC to DC.
Component Suggestion:
- Bridge Rectifier IC: KBPC5010 or equivalent, capable of handling the required current and voltage.
### b. Voltage Drops
- Diode Drops: Each diode in the bridge typically drops about 0.7V.
- Total Drop: 1.4V (since two diodes conduct in series during each half-cycle).
### c. Reservoir Capacitors
To smooth the rectified DC:
- Per Rail: 1000µF electrolytic capacitor.
Recommendation:
- Voltage Rating: At least 25V (preferably 35V for safety).
### d. Post-Regulation Filtering
After the linear regulators:
- Per Rail: 100µF electrolytic capacitor.
- High-Frequency Decoupling: 100nF ceramic capacitor across each electrolytic.
Placement:
- Close to Regulators: To effectively filter out high-frequency noise.
---
## 4. Voltage Regulation
### a. Selecting Regulators
- Positive Rail: LM7815 (15V regulator)
- Negative Rail: LM7915 (-15V regulator)
### b. Voltage Drop Across Regulators
- Input to LM7815/LM7915: ~19.8V DC
- Output: ±15V DC
- Voltage Drop: ~4.8V per regulator
### c. Power Dissipation in Regulators
\[
P_{\text{reg}} = V_{\text{drop}} \times I_{\text{out}}
\]
Assuming 1A per rail:
\[
P_{\text{reg}} = 4.8V \times 1A = 4.8W
\]
Recommendation:
- Heatsinks: Attach adequate heatsinks to both LM7815 and LM7915 to dissipate heat effectively.
- Thermal Management: Consider using thermal compound and ensure good airflow around the regulators.
---
## 5. Capacitor Selection and Sizing
### a. Reservoir Capacitors (Before Regulators)
- Value: 1000µF per rail
- Voltage Rating: ≥25V
### b. Post-Regulator Capacitors
- Value: 100µF per rail
- Voltage Rating: ≥25V
### c. High-Frequency Decoupling
- Value: 100nF ceramic capacitors
- Placement: Across each electrolytic capacitor, close to the regulator pins.
### d. Calculating Filter Capacitor Size
To ensure minimal ripple voltage, the capacitor size should be sufficient based on the load current and the desired ripple.
\[
\Delta V = \frac{I}{f \times C}
\]
Where:
- \( \Delta V \) = Ripple voltage
- \( I \) = Load current
- \( f \) = Frequency (for full-wave rectification, \( f = 120Hz \) in a 60Hz system)
- \( C \) = Capacitance
Example Calculation:
Assuming:
- I = 1A
- f = 120Hz
- C = 1000µF
\[
\Delta V = \frac{1}{120 \times 1000 \times 10^{-6}} \approx 8.3V
\]
This high ripple is unacceptable. Therefore, increasing capacitor size or reducing load current is necessary.
Adjusted Calculation for Acceptable Ripple (~1V):
\[
C = \frac{I}{f \times \Delta V} = \frac{1}{120 \times 1} \approx 8333µF
\]
Recommendation:
- Increase Reservoir Capacitors: Use 10,000µF per rail to reduce ripple effectively.
---
## 6. Transformer VA Rating Correction
Your initial estimation suggested a 5VA transformer, but based on detailed calculations, a 10VA transformer is more appropriate. This ensures:
- Adequate Headroom: Prevents voltage sag under load.
- Longevity: Reduces stress on transformer components.
- Thermal Management: Minimizes overheating risks.
---
## 7. Practical Implementation Steps
### a. Components List
1. Transformer:
- Dual 15VAC center-tapped, 10VA
2. Rectification:
- Bridge Rectifier (e.g., KBPC5010)
- Four 1N4007 diodes (if building your own bridge)
3. Filtering:
- Electrolytic Capacitors:
- 10,000µF, 25V per rail
- Post-Regulator:
- 100µF, 25V per rail
- Decoupling:
- 100nF ceramic capacitors per rail
4. Voltage Regulators:
- LM7815 for +15V
- LM7915 for -15V
- Heatsinks for both regulators
5. Additional Components:
- Fuse: 1A slow-blow for input protection
- NTC Thermistor: For inrush current limiting
- PCB or Breadboard: For assembly
- Wire and Connectors: As needed
### b. Assembly Steps
1. Transformer Setup:
- Connect the primary to mains (ensure correct and safe wiring).
- Identify the center-tap as the ground (0V).
2. Bridge Rectifier:
- Connect the secondary windings to the bridge rectifier.
- Ensure correct polarity: AC inputs to ~, + and - outputs appropriately.
3. Reservoir Filtering:
- Connect 10,000µF capacitors to each rail post-rectification.
- Observe correct polarity to prevent capacitor damage.
4. Voltage Regulation:
- Connect LM7815 to the +V rail.
- Connect LM7915 to the -V rail.
- Attach heatsinks to both regulators.
- Add 100µF capacitors post-regulation for stability.
5. Decoupling:
- Place 100nF capacitors across each electrolytic capacitor pair near the regulator inputs and outputs.
6. Protection:
- Insert a 1A slow-blow fuse on the input side to protect against overcurrent.
- Add an NTC thermistor in series with the transformer primary for inrush current limiting.
7. Testing:
- Before Powering Up:
- Double-check all connections.
- Verify no shorts exist between rails.
- Power Up:
- Use a variac or a current-limited power supply for initial testing.
- Measure DC voltages to ensure correct ±15V outputs.
- Check for excessive heat on regulators.
---
## 8. Fine-Tuning and Validation
### a. Ripple Voltage Measurement
Use an oscilloscope to check the ripple voltage on the output rails. Aim for:
- ±15V Rail: Ripple < 1V RMS
If ripple is higher, consider increasing the reservoir capacitance or improving filtering.
### b. Load Testing
- Apply Load: Connect a dummy load (e.g., power resistors) equivalent to your headphone impedance.
- Monitor Voltages: Ensure regulators maintain stable outputs under load.
- Thermal Monitoring: Check that regulators do not overheat.
### c. Audio Performance Testing
- Connect Headphones: Start at low volume.
- Listen for Distortion: Ensure audio clarity without unwanted noise.
- Adjustments: If needed, tweak filtering or regulator configurations to optimize sound quality.
---
## 9. Safety Considerations
- Isolation: Ensure the transformer provides adequate isolation from mains to prevent electrical hazards.
- Component Ratings: All capacitors and regulators should exceed the maximum expected voltages.
- Heat Management: Proper heatsinking prevents thermal runaway and prolongs component life.
- Fusing: Protects against short circuits and overcurrent scenarios.
- Ventilation: Ensure the power supply area is well-ventilated to dissipate heat.
---
## 10. Summary of Adjusted Component Values and Recommendations
1. Transformer:
- Voltage: Dual 15VAC center-tapped
- VA Rating: 10VA (Recommended: 10VA or higher)
2. Rectification:
- Bridge Rectifier: KBPC5010 or equivalent
- Diodes (if separate): Four 1N4007
3. Filtering:
- Reservoir Capacitors: 10,000µF, 25V per rail
- Post-Regulator Capacitors: 100µF, 25V per rail
- Decoupling Capacitors: 100nF ceramic per rail
4. Voltage Regulators:
- Positive Rail: LM7815 with heatsink
- Negative Rail: LM7915 with heatsink
5. Protection:
- Fuse: 1A slow-blow
- NTC Thermistor: For inrush current limiting
6. Additional:
- High-Frequency Caps: 100nF ceramics across electrolytics
- Heat Sinks: Appropriately sized for LM7815 and LM7915
---
## 11. Additional Tips
- Regulator Placement: Keep regulators away from high-current paths to minimize noise coupling.
- PCB Layout: If designing a PCB, separate analog and power grounds to reduce interference.
- Component Quality: Use low Equivalent Series Resistance (ESR) capacitors for better filtering.
- Modularity: Design the power supply in modules for easier troubleshooting and upgrades.
---
## 12. Final Validation and Testing
After assembling the power supply:
1. Initial Power-On:
- Use a variac or current-limited source to gently apply power.
- Verify DC voltages before connecting to the amplifier.
2. Load Testing:
- Connect a resistor (e.g., 15Ω) to simulate the headphone load.
- Monitor voltage stability and temperature.
3. Audio Testing:
- Connect the headphones.
- Play audio at various volume levels.
- Listen for clarity, absence of hiss or hum, and overall sound quality.
4. Iterative Refinement:
- Based on testing, adjust capacitor sizes or regulator configurations if necessary.
- Ensure all components operate within safe thermal and electrical limits.
---
By following this refined approach, you can create a robust, low-voltage power supply tailored for your headphone amplifier needs. This ensures that your HE1000 Stealth headphones receive clean, stable power, preserving their pristine and tubey sound characteristics while maintaining low total harmonic distortion and noise.
Feel free to reach out if you need further assistance or clarification on any of these steps!
So we are going with separate transformers for high and low voltages, it seems. I hope there is room in the whammy chassis (80mm high x 210mm wide x 230mm deep). If we need a different chassis, that's fine.
Well, that's something I've been having trouble with. I've been looking for a toroidal tranformer that has secondaries for ~150-200V for the tube, 6.3V for the heater, and 24V for the mosfets (when I was thinking I'd use irlz44n). Now, I will look for 150V, 6.3V and 15V? I haven't found anything on antek's website that looks specific for a tube hybrid.